The applications of ultrasound in surgical practice are widespread (Table 1) 115. The great majority of examinations are performed either in obstetric patients or are undertaken to study the abdomen. It is not within the scope of this chapter to consider in detail obstetric ultrasound. The ability to obtain images of the fetus and placenta without the use of ionizing radiation has greatly enhanced obstetric practice, and in the United Kingdom most pregnancies will be scanned at some stage during their course. It is possible to assess accurately gestational age, placental position, and the presence of many different fetal anomalies as well as assessing different parameters in the growth-retarded fetus.

The use of non-ionizing radiation makes ultrasound safe; it is also relatively inexpensive and can be performed at the patient's bedside. A number of different organs can be viewed at the same examination and in a short space of time. Ultrasound is therefore widely practised by surgeons, cardiologists, and neonatologists as well as radiologists.

Ultrasound has limitations, in that the interpretation of images obtained is dependent on the skill of the operator and image quality is degraded by obesity and by the presence of bowel gas.

Patients with a palpable or suspected abdominal mass are investigated in a variety of ways. If there is a strong clinical suspicion that the mass originates in the colon, for example, then a barium enema examination may be requested as the initial examination. Since it is unusual for the organ of origin of a mass to be known, patients are frequently referred initially for examination by either ultrasound or computed tomography (CT). Both CT and ultrasound have a positive predictive value of between 95 per cent and 100 per cent in determining the presence or absence of an abdominal mass, and the predicted organ of origin of a mass is correct in 90 per cent of patients examined with either technique. CT is slightly more sensitive in its ability to define the exact nature of a mass. Since there is little to choose between the two techniques, the ease, lower cost, and lack of exposure to ionizing radiation, make ultrasound the method of choice.

Ultrasound examination of the right hypochondrium includes a detailed assessment of liver texture, the bile ducts, gallbladder, and right kidney.

Ultrasound has replaced the oral cholecystogram for the detection of gallstones, not so much because of its greater sensitivity but rather because it enables examination of other structures at the same time. There is little difference in the sensitivity of the two examinations in the detection of gallbladder pathology. It is not possible to state the actual sensitivity, since a negative examination by either technique is often the end point and thus false negative diagnoses are only rarely detected.

In acute cholecystitis the gallbladder is distended and thick-walled, and stones will usually be seen (Fig. 1) 161. Pericholecystic inflammatory change may also be visible. As it is possible to localize the position of the gallbladder precisely, direct pressure on this area by the transducer will elicit a positive ultrasound Murphy's sign.

The ability of ultrasound to detect bile duct dilatation, in the intrahepatic bile duct especially, approaches 100 per cent and it is the primary investigation in patients with suspected obstruction (Fig. 2) 162. If biliary dilatation is detected, the level of obstruction needs to be defined: this is possible in 95 per cent of cases, while the cause of the obstruction can be determined in 85 per cent of patients. The sensitivity varies depending on the obstructing lesion (Table 2) 116. Direct visualization of tumours of the ampulla of Vater is rare, as is the precise diagnosis of sclerosing cholangitis.

Once the diagnosis of biliary obstruction has been made it may be appropriate to perform a CT scan if ultrasound has failed to demonstrate the cause of obstruction or if surgery is contemplated.

Although ultrasound has a major role in investigating diffuse hepatocellular disease, in surgical practice it is most commonly used to determine the presence of focal liver lesions (Fig. 3) 163. It is able to differentiate between solid and cystic lesions, and if there is any doubt about the nature of a mass guided aspiration can be performed to determine whether the lesion is a benign or malignant tumour or whether it is an abscess or cyst. The use of ultrasound at the time of surgery, with a sterile transducer applied directly to the liver, has led to a reappraisal of the sensitivity of preoperative investigations in the detection of liver tumours. Intraoperative ultrasound is currently the most sensitive method of detecting liver tumours: 40 per cent of the lesions detected by this method are not palpable at surgery. Preoperative assessment with either ultrasound or conventional CT has a sensitivity of 60 to 70 per cent in the detection of liver tumours. In some centres magnetic resonance imaging is preferred as it has the ability to differentiate between benign capillary haemangiomas and metastases.

Ultrasound fails to produce a good image of part or all of the pancreas in 25 per cent of patients, due to the presence of overlying bowel gas. In the majority of patients in whom the pancreas is clearly seen detection of abnormalities is sensitive. If the pancreas is poorly visualized and if surgery is contemplated, or in acute pancreatitis, CT should also be performed.

Focal masses of altered echogenicity are usually due to adenocarcinoma. Occasionally guided percutaneous biopsy is required to allow differentiation of such masses from focal areas of pancreatitis or rare benign tumours. Endocrine tumours of the pancreas are often small and difficult to detect preoperatively by any technique: intraoperative ultrasound visualizes these with a sensitivity approaching 100 per cent.

During an acute attack of pancreatitis the coexisting ileus may obscure the pancreas. CT is the preferred investigation, particularly as this allows an assessment of pancreatic necrosis. Ultrasound can be used to evaluate the gallbladder and biliary tree and to monitor pancreatic fluid collections.

A combination of endoscopic retrograde pancreatography and ultrasound is the most sensitive method for detecting chronic pancreatitis. In a few patients CT will provide additional information. There is, however, poor correlation between the findings on any imaging study and the degree of pancreatic dysfunction other than in patients with the most severe disease.

Ultrasound reliably demonstrates the shape, size, and echogenicity of the kidneys and the nature of mass lesions. It will also demonstrate renal tract dilatation. The normal non-dilated ureter is not visualized. The bladder and prostate are well seen with both transabdominal and transrectal ultrasound. Ultrasound is therefore the primary imaging modality in renal failure, urinary tract infection, and in the assessment of bladder outflow obstruction.

Ultrasound is a highly sensitive means for the detection of renal tract dilatation. While dilatation usually indicates obstruction this is not always so, and obstruction in its early phases does not always produce dilatation. Ultrasound provides no functional information and the level of obstruction may be difficult to discern. For these reasons it will often need to be followed by urography, antegrade pyelography, or renal scintigraphy. Urinary tract calcification, especially in the ureters, may not always be detected by ultrasound and a plain abdominal radiograph should usually accompany any ultrasound examination of the renal tract. Patients with symptoms suggestive of ureteric calculi should undergo urography rather than ultrasound examination.

Ultrasound is more sensitive and more specific than urography in the detection of renal mass lesions other than tumours of the renal pelvis, which are also more reliably diagnosed by CT. Ultrasound may be used to determine whether mass lesions detected by urography are cystic or solid (Fig. 4) 164. Tumours of the ureter are poorly detected by ultrasound unless they are causing hydronephrosis. Tumours of the bladder are better detected by transabdominal ultrasound than urography, their respective sensitivities being 77 per cent and 61 per cent, respectively. The combination of transabdominal and transrectal ultrasound increases the sensitivity to 95 per cent.

The initial investigation in patients with haematuria is usually urography. If cystoscopy is not to be performed then the urogram should be accompanied by an ultrasound scan. If ultrasound is used as the first investigation there is a risk that urothelial tumours of the renal pelvis and ureter will be missed. These tumours are however very much less common than renal cell carcinoma and bladder tumours.

Transabdominal ultrasound permits a two-dimensional volumetric assessment of the degree of intravesical enlargement of the prostate, as well as assessment of residual urine volume. It is therefore preferred to urography in patients with bladder outflow obstruction. Transrectal ultrasound gives improved resolution of images of the prostate itself, allowing breaches of the capsule and adjacent lymph node involvement by tumours to be detected.

The uterus and ovaries may be visualized by both transabdominal and transvaginal scanning. Transvaginal ultrasound has the advantage that the transducer is nearer the midline pelvic structures and may therefore be of a higher frequency, which improves the resolution of the image. This method of scanning is preferred in the early detection of pregnancy and in the assessment of abortion and ectopic pregnancy. The higher frequency limits the depth of penetration of the ultrasound wave and more laterally situated pelvic masses are often better assessed by transabdominal scanning.

The role of ultrasound in the assessment of a pelvic mass is to determine the nature and size of the mass as well as its organ of origin. Ultrasound reliably differentiates between solid and cystic lesions in the pelvis. In general terms most ovarian lesions are cystic and most uterine lesions are solid; there is some overlap, however, and the organ of origin is then inferred by visualizing either the normal ovaries or the normal uterus and excluding these as the cause for the mass. A very large mass may obscure normal structures, making the organ of origin more difficult to predict. Features suggestive of malignancy in ovarian lesions are a multiloculated appearance and the presence of solid elements within the mass (Fig. 5) 165. Ultrasound is of limited value in the evaluation of malignancy of the uterine body or cervix although uterine fibroids are well visualized. Ultrasound will also permit differentiation between masses of gynaecological origin and other pelvic masses, such as bowel lesions.

Recently much attention has been given to the possibility of using ultrasound to screen for ovarian cancer. Using transabdominal ultrasound imaging, the odds that a positive result on screening indicates the presence of a primary ovarian cancer are 1 in 67. The sensitivity can be improved by using transvaginal colour flow Doppler ultrasound, which reduces the number of false positive results. Colour flow Doppler ultrasound machines cost two to three times as much as conventional ultrasound equipment, however.

In young women it can be difficult to differentiate between appendicitis, pelvic inflammatory disease, and complications of ovarian cysts. In this clinical setting ultrasound is of value if a positive diagnosis of one of these causes for the pain can be made.

The advent of transvaginal ultrasound has increased the likelihood of making a positive diagnosis of ectopic pregnancy. It also permits the earlier diagnosis of an intrauterine pregnancy, which effectively eliminates the diagnosis of ectopic pregnancy as the two only very rarely coexist. Some form of pelvic abnormality will be seen in three-quarters of patients with an ectopic pregnancy, although the ectopic gestation sac and embryo will be seen in only 40 per cent. As 25 per cent of patients with ectopic pregnancy will have a virtually normal scan it is imperative that the scan is always performed in conjunction with an accurate measurement of human chorionic gonadotrophin.

Not only does ultrasound allow an assessment of the morphology of the pelvic organs but it also permits monitoring of cyclical ovarian and endometrial changes. For the purposes of in-vitro fertilization, transvaginal ultrasound allows accurate needle aspiration of ovarian follicles to harvest ova.

With the exception of the diagnosis of pyloric stenosis in infancy, ultrasound does not have a primary role in the investigation of the gastrointestinal tract. However tumours of the gastrointestinal tract will often be obvious on ultrasound, although it may be difficult to differentiate these from inflammatory masses such as diverticular masses. Bowel wall thickening and stricture formation is also often apparent in conditions such as Crohn's disease. Ultrasound has a sensitivity of more than 80 per cent and a specificity of 95 per cent in the diagnosis of acute appendicitis without perforation and it has been advocated that it should be used routinely in an attempt to reduce the negative appendicectomy rate. The sensitivity is much lower in patients with perforated appendices, although it has been argued that this does not matter as the need for surgery will be obvious. This fact, and the fact that ultrasound will not detect up to 20 per cent of cases of acute appendicitis, means that it should be used judiciously. It should probably be confined to patients with equivocal clinical findings and young women, in whom it will exclude a gynaecological cause for the pain. Surgeons should be aware of the high specificity of a positive diagnosis but must equally be aware that a negative result must be treated with caution.

Much of the recent development in ultrasound has related to the technological advances that now permit the incorporation of ultrasound transducers into endoscopes and the development of intraluminal transducers. Using transducers attached to upper gastrointestinal tract endoscopes it is now possible to obtain images of the oesophagus and heart with a transducer within the oesophagus and to visualize the pancreas, duodenum, and common bile duct with a transducer in the stomach or duodenum. If the endoscope is able to pass through an oesophageal tumour, ultrasound is more sensitive than CT for local staging. It is able to define extension through the layers of the oesophagus, extension to adjacent organs, and spread to local lymph nodes. If the tumour cannot be passed then CT is superior for the detection of mediastinal extension.

Ultrasound examination of the rectum and colon can be performed using either a rectal transducer or a transducer incorporated into a colonoscope. Correlation of the ultrasound results with histological analysis of the resected specimen according to the 1987 TNM classification gives overall accuracy rates of 81 per cent for rectal tumours and 93 per cent for colonic tumours. Some problems have been found with T2 tumours, which may be accompanied by peritumoural inflammation and abscess formation. The possibility of false positive diagnosis of lymph node metastasis is a problem.

Distant metastases within the abdomen can only be detected by abdominal ultrasound or CT.

Ultrasound, CT, and nuclear medicine techniques may be used to detect intra-abdominal abscesses. No technique is capable of specifically determining whether an identified fluid collection is infected or not, and percutaneous aspiration of the fluid should always be performed. Many comparative studies of the sensitivity of the three techniques have been performed: those for ultrasound range from 60 to 100 per cent, for CT from 78 to 100 per cent, and for nuclear medicine techniques 75 to 100 per cent. Abscesses in the suphrenic spaces, the right upper quadrant, perirenal areas, and the midline of the pelvis are well seen with ultrasound. CT has the advantage of not being affected by open wounds, dressings, and bowel gas, unlike ultrasound, and CT is better able to image the mesentery and retroperitoneum. Ultrasound should be used initially to look for an abscess. If this is inconclusive then CT will overcome most of the problems that make the ultrasound scan unsatisfactory. Nuclear medicine techniques using either gallium-67 or labelled white cell scanning should be reserved for those patients in whom intra-abdominal sepsis is strongly suspected but in who a definitive diagnosis cannot be reached by the other two techniques.

Percutaneous fluid drainage and biopsy procedures can be guided by conventional fluoroscopic screening, CT, or ultrasound. The method chosen will be influenced by the site within the abdomen and the individual preference of the operator. Masses or abscesses within the retroperitoneum, mesentery, or pelvis are often better approached using computed tomography as this provides better visualization of, and therefore avoidance of, the bowel. Superficial lesions, lesions in the liver, and collections in the suphrenic spaces can usually be approached with ease using ultrasound.

The ability to drain abscesses percutaneously has totally changed the management of such patients. Surgery with its greater morbidity, mortality, and cost can often be avoided, or at least deferred until the patient's general condition improves. Success rates of 75 to 85 per cent for percutaneous abscess drainage procedures are common.

Percutaneous biopsy may be performed using either a fine needle to obtain specimens for cytological examination or using larger needles for histological specimens. If accurate cytology is available then fine needle aspiration biopsy is preferred because of its lower complication rate. Large surveys of the complication rate of fine needle biopsy in Europe and the United States reveal mortality rates between 0.006 and 0.031 per cent and rates of tumour seeding along the needle tract of between 0.003 and 0.009 per cent. Most deaths have been due to haemorrhage following liver biopsy or, less commonly, pancreatitis after pancreatic biopsy.

Many other procedures including nephrostomy, biliary drainage, and percutaneous pancreatography can be performed using ultrasound guidance.

The ability to differentiate between cystic and solid areas allows assessment of focal or diffuse thyroid enlargement. Most lesions within the thyroid are palpable, and the availability of fine needle aspiration biopsy for cytology means that there is often no value in imaging thyroid nodules, particularly as ultrasound cannot reliably separate benign from malignant lesions.

Ultrasound will frequently detect enlargement of the parathyroid glands if they are in a conventional site but it is unable to detect hyperplasia of the glands. In the search for a parathyroid adenoma it should therefore be the first investigation (Fig. 6) 166. In rare instances ultrasound can be used to guide aspiration of impalpable thyroid and parathyroid lesions. Occasionally ablation of a parathyroid adenoma by ultrasound guided alcohol injection has been performed in patients unfit for surgery. Any mass in the neck or region of the salivary glands can be evaluated with ultrasound and its nature and organ of origin characterized.

Ultrasound is highly sensitive in the detection of non-palpable testicular tumours and may therefore be used to exclude the presence of a tumour in patients with non-specific scrotal symptomatology. Tumour echo patterns vary and classification of tumour type is not possible (Fig. 7) 167. Some benign testicular conditions such as orchitis may mimic tumours. If a tumour is detected then abdominal ultrasound should be performed to detect lymph node metastases, although actual staging is performed using CT.

Epididymo-orchitis may produce characteristic thickening of the epididymis which may aid in its differentiation from testicular torsion. Ultrasound imaging of the testis is unreliable in diagnosing torsion, and radionuclide imaging is currently the investigation of choice, with a reported sensitivity of 90 per cent. Continuous wave Doppler stethoscopes and conventional duplex Doppler ultrasound have given unreliable results. Recent studies using colour flow Doppler ultrasound report sensitivities of 86 and 100 per cent and specificity of 100 per cent in the detection of testicular torsion, and this may well be the method of the future.

Trauma to the testis, such as testicular rupture and testicular and paratesticular haematoma are well demonstrated.

The value of duplex ultrasound in the investigation of penile erectile dysfunction is currently being assessed. Initial studies suggest that this may be a valuable technique as it is possible to assess peak systolic velocity in the cavernosal arteries as well as diminished flow in diastole as indicators of the adequacy of the veno-occlusive mechanism.

Although there are those who support the use of ultrasound as a primary imaging technique, it is most often used to clarify the cystic or solid nature of a mass detected on X-ray mammography or in breasts that appear particularly dense on mammography. Ultrasound should be used initially in the assessment of inflammatory breast disease as abscesses can be localized, and percutaneous drainage can be performed. It may also be used to evaluate breast masses in young patients.

One of the earliest applications of ultrasound in the evaluation of musculoskeletal system, and still one of the most commonly performed examinations, is the examination of painful hips, particularly in children. This is an extremely simple and accurate technique for detecting fluid within the joint and, if appropriate, for guiding joint aspiration.

Dynamic examination of the hip in infancy using ultrasound allows detection of congenital dislocation at a time when the femoral head is not ossified and is therefore not visible on the plain radiograph. Other applications of ultrasound to the musculoskeletal system are the detection of non-radio-opaque foreign bodies, the evaluation of rotator cuff and Achilles' tendon injuries and the assessment of soft tissue masses and haematomas.

Duplex ultrasound imaging of the extracranial carotid arteries is now widely performed in patients who have experienced a transient ischaemic attack. It may be used to determine which patients go on to angiography and possible surgery. In experienced hands ultrasound has a 92 to 95 per cent accuracy as compared with conventional arteriography in the analysis of extracranial carotid artery disease. If the duplex examination demonstrates a completely normal vessel in a patient with a classical transient ischaemic attack or stroke the cause is unlikely to lie in the extracranial carotid artery. Duplex ultrasound may be used to evaluate the arterial and venous systems of the arms and legs, and in many centres it has replaced venography for the diagnosis of lower limb venous thrombosis. Thrombosis in the popliteal, femoral, and proximal external iliac veins is readily confirmed or excluded, although thrombosis confined to the deep veins of the calf is less easy to demonstrate. Venography is therefore preferred if anticoagulation is to be used for thrombosis confined to the deep calf veins.

The intra-abdominal vasculature, particularly the hepatic and portal venous systems and the renal vasculature, may be readily evaluated with duplex and colour flow Doppler ultrasound. The altered dynamics of the portal venous system in portal hypertension and obstruction to the hepatic veins are all well demonstrated. Doppler evaluation in possible renovascular hypertension has been disappointing but the assessment of the renal venous system and the characterization of renal perfusion in renal transplants is more promising. Not only does duplex ultrasound of the transplanted kidney give an indication of rejection but it will also detect obstruction of the kidney and the presence of perirenal fluid collections such as lymphoceles and urinomas.

With very few exceptions ultrasound demonstrates the presence or absence of abdominal aortic aneurysms and can be used for subsequent monitoring of their size. It is less reliable than angiography and computed tomography in establishing the relationship to the renal arteries and less reliable than computed tomography in detecting a leaking aneurysm.

General

Bourne T, Campbell S, Steer C, Whitehead MI, Collins WP. Transvaginal colour flow imaging: a possible new screening technique for ovarian cancer. Br Med J, 1989; 299: 1367–70.

Campbell S, Bhan V, Whitehead MI, Collins WP. Transabdominal ultrasound screening for early ovarian cancer. Br Med J, 1989; 299: 1363–7.

Cooperberg PL, Gibney RG. Imaging of the gallbladder. Radiology, 1987; 163: 605–13.

Ferrucci JT. Liver tumour imaging: current concepts. Am J Roentgenol, 1990; 155: 473–84.

Freeny PC. Radiology of the pancreas: two decades of progress in imaging and intervention. Am J Roentgenol, 1988; 150: 975–81.

Lindsell DRM. Ultrasound imaging of pancreas and biliary tract. Lancet, 1990; 335: 390–3.

Puylaert JBCM, et al. A prospective study of ultrasonography in the diagnosis of appendicitis. N Engl J Med, 1987; 317: 666–9.

Spencer J, Lindsell D, Mastorakou I. Ultrasonography compared with intravenous urography in the investigation of adults with haematuria. Br Med J, 1990; 301: 1074–6.

Thorsen MK, et al. Diagnosis of ectopic pregnancy: endovaginal vs transabdominal sonography. Am J Roentgenol, 1990; 155: 307–10.

Webb JA. Ultrasonography in the diagnosis of renal obstruction. Br Med J, 1990; 301: 944–6.

Specific

1. Hepatobiliary, pancreatic, and gastrointestinal ultrasound.

Kurtz AB, Goldberg BB. Gastrointestinal ultrasonography. Clinics in diagnostic ultrasound 23. New York: Churchill, Livingstone, 1988.

2. Genitourinary ultrasound.

Hricak H. Genitourinary ultrasound: Clinics in Diagnostic Ultrasound 18. New York: Churchill Livingstone, 1986.

3. Gynaecological ultrasound.

Steel WB, Cochrane WJ. Gynecologic ultrasound: Clinics in Diagnostic Ultrasound 15. New York: Churchill Livingstone, 1984.

4. Endoscopic and intraoperative ultrasound.

Rifkin MD. Intraoperative and endoscopic ultrasonography: Clinics in Diagnostic Ultrasound 22. New York: Churchill Livingstone, 1987.

5. Vascular and Doppler ultrasound.

Taylor KJW, Strandness DE. Duplex Doppler ultrasound: Clinics in Diagnostic Ultrasound 26. New York: Churchill Livingstone, 1990.