2D sonography may show thickened bowel wall, pneumatosis intestinalis, portal venous gas and gas in the periphery of the liver, free abdominal fluid and free intra-abdominal gas (Silva et al. 2007). All above-mentioned sonographic signs however are not pathognomonic for NEC. The more of these signs can be found, the more probable NEC is.

Similar to other inflammatory bowel diseases thickening of the bowel wall can be seen. Cross section of the affected bowel wall shows a cockade with a hyperechoic centre and a surrounding hypoechoic ring with decreased echogenicity (Fig. 9.11a). Longitudinal sections through the affected bowel show a tubular structure with an echogenic lumen surrounded by two hypoechoic linear structures, the thickened bowel wall (Fig. 9.11b).

2D sonography can clearly depict intramural gas bubbles which appear as punctate focal areas of increased echogenicity within the thickened bowel wall (Fig. 9.12a, b). They represent pneumatosis intestinalis (Fig. 9.12d). Large quantities of intramural gas cause multiple dots of increased echogenicity in the anterior wall of the bowel which prevent sound transmission to the posterior wall, which can no longer be displayed (Fig. 9.12c). The gas can get access to the superior and inferior mesenteric veins. From here, it is transported to the portal vein and into the liver. In these cases, gas bubbles can be seen as moving areas of focal increased echogenicity within the portal vein (Fig. 9.13a, b). The gas bubbles gather in the capillary bed of the liver where they appear as multiple areas of increased echogenicity (Fig. 9.13c).

Pulsed Doppler recording of the flow within the portal vein shows multiple artefacts in the form of irregular peaks within the Doppler spectrum caused by the gas bubbles (Fig. 9.13d).

Free abdominal fluid can be found in NEC too. It can be shown either in the lower abdomen behind the bladder or in the Morrison pouch between the liver and the right kidney (Fig. 9.14a). Large amounts of fluid can also be seen in longitudinal sections through the right or left flank (Fig. 9.14b). Free abdominal fluid without internal echoes is not a sign of intestinal perforation (Fig. 9.14a, b). The presence of intraperitoneal fluid at US, particularly if it is complex or echogenic, is suggestive of bowel perforation (Fig. 9.14c). Free abdominal fluid is much more accurately depicted with US than by abdominal radiography. This finding, however, is nonspecific and may also be seen in intraperitoneal haemorrhage, with other causes of peritonitis, or in the presence of peritoneal dialysis catheters.

If the free abdominal fluid contains moving internal echoes, it can be displayed with colour Doppler (Fig. 9.20a, b). Due to the respiratory excursions of the diaphragm, the fluid moves to and fro (Fig. 9.20a, b).

Free abdominal gas is traditionally diagnosed with abdominal radiography performed with a horizontal beam but may also be detected with US. Free abdominal gas is suspicious of intestinal perforation. It can be seen below the costal margin in the region of the liver or spleen (Fig. 9.15a). Free intra-abdominal gas causes multiple reverberation echoes (Fig. 9.15a). In the case of intestinal perforation, these organs disappear during respiration behind a curtain of multiple reverberation echoes caused by free abdominal gas (Fig. 9.15b).

Doppler sonographic flow measurements in the abdominal arteries of patients with NEC have shown increased flow velocities in the superior mesenteric artery (SMA) in comparison with healthy infants (Fig. 9.16) (Kempley and Gamsu 1992; Deeg et al. 1993). Kempley and Gamsu measured the mean blood flow velocities in the SMA of infants with confirmed NEC, suspected NEC in comparison with a healthy control group. SMA mean blood flow velocities were significantly higher in infants with confirmed necrotising enterocolitis (36.5 cm/s) than in unfed healthy controls (20.4 cm/s) or infants with unconfirmed enterocolitis (19.6 cm/s) (Kempley and Gamsu 1992). Kempley and Gamsu found no difference in the flow velocities in the coeliac trunk and the anterior cerebral artery in the different groups.

We found a significant increase of the peak and time average flow velocities in the SMA in comparison with an age- and weight-matched healthy control group (Fig. 9.16b) (Deeg et al. 1993).

In patients with NEC, peak systolic flow velocities were 119 ± 58 cm/s (control, 68 ± 21 cm/s), and time average velocities were 28 ± 13 cm/s (control, 13 ± 6 cm/s). The end-systolic and end-diastolic flow velocities were not significantly different in both groups (Fig. 9.17b) (Deeg et al. 1993) (Table 9.1).

In contrast to Kempley and Gamsu, we also found a significant increase of the flow velocities in the coeliac trunk (Fig. 9.17) (Deeg et al. 1993). In patients with NEC, peak systolic flow velocities in the coeliac trunk were 124 ± 40 cm/s (control, 73 ± 11 cm/s), and time average velocities were 40 ± 15 cm/s (control, 20 ± 5 cm/s) (Fig. 9.17c) (Deeg et al. 1993). The end-systolic and end-diastolic flow velocities were not different in both groups (Table 9.2).

The interpretation of the increased peak systolic and time average flow velocities in patients with NEC is difficult. It is probably caused by inflammation of the affected bowel. The increase of the flow velocities can be caused either by an increase of the volume flow or by a decrease of the vessel diameter. An increase of the volume flow in the very sick premature infants with problems in maintaining a stable blood pressure is very improbable. An increased bowel perfusion would also increase the end-systolic and end-diastolic flow velocities – increases which could not be found in our patients. The most probable explanation for the increased blood flow velocities within the mesenteric arteries is vasoconstriction. The lack of an increase of the diastolic flow velocities (end systolic and end diastolic) and a simultaneous increase of the resistive index in the coeliac trunk (from 0,72 ± 0,10 in control infants to 0,80 ± 0,09 in patients with NEC) constitute a strong argument for the vasoconstriction theory (Deeg et al. 1993). According to the Bernoulli equation, flow velocities increase if the vessel diameter decreases. A further reduction of the vessel diameter could eventually cause a decrease of the flow velocities. Due to the pulsatile blood flow in the arterial system, systolic and diastolic flow alterations can differ: This may cause a decrease of the diastolic flow although the systolic flow further increases.

Other authors however favour the hyperperfusion theory (Kempley and Gamsu 1992).

If a baby is operated upon, the serosal surface of affected bowel is often noticed to be inflamed; the increased SMA and TC velocities in necrotising enterocolitis could therefore result from an inflammatory response in the remaining ‘healthy’ bowel (Kempley and Gamsu 1992). Kempley and Gamsu states that flow velocities in the mesenteric arteries may be increased, due to the inflammatory process involving the bowel and gut. This may increase volume flow and total bowel perfusion although the affected bowel is hypoperfused.

Flow measurements in the portal vein may demonstrate flow alterations caused by intraportal gas bubbles (Fig. 9.13d). They cause a characteristic audio signal and small peaks within the otherwise smooth flow spectrum of the portal vein (Fig. 9.13d).

As mentioned earlier, several cardiovascular abnormalities are associated with an increased risk of developing NEC, especially leakages of the aortic Windkessel and obstructions of the left heart. Leakages of the aortic Windkessel cause a very low diastolic flow (Fig. 9.18a, b), whereas left heart obstructions lead to a reduced systolic flow (Fig. 9.18c). In both cases, the time average flow velocity is reduced which causes a dramatic decrease of volume flow and bowel perfusion.

Faingold et al. have shown that with NEC, a variety of patterns of bowel wall perfusion may be depicted with colour Doppler US (Faingold et al. 2005). They state that the absence of bowel wall perfusion at colour Doppler US is more sensitive and specific than the presence of free air at abdominal radiography in the detection of necrotic bowel in NEC (Faingold et al. 2005).

Faingold et al. developed a standardised colour Doppler US protocol that detected bowel wall perfusion accurately and reproducibly in healthy infants (Faingold et al. 2005). Colour Doppler parameters included the lowest possible pulse repetition frequency without aliasing, a low wall filter, and the highest Doppler gain settings without flash artefacts. Velocity was set at 0.029–0.11 m/s. Mural perfusion was evaluated by counting the number of colour Doppler signals as dots and lines per square centimetre. Colour Doppler flow was determined to be present when colour Doppler signals were reproducible or confirmed with pulsed Doppler waveforms (Faingold et al. 2005). In patients with suspected NEC, bowel loops were interrogated for the presence or absence of mural blood flow and compared with adjacent loops. When bowel loops without mural blood flow were identified, the settings were changed to more sensitive settings in an attempt to detect the slowest possible velocities down to 0.029 m/s, below which only flash artefacts were obtained. A variety of patterns of increased bowel wall perfusion could be seen in the patients with Bell stages I and II NEC. Increased bowel wall perfusion was depicted as a marked increase in colour Doppler signals in the bowel wall and/or mesentery (Fig. 9.19). It was interpreted as reflecting marked hyperaemia due to vasodilatation (Faingold et al. 2005). In some involved loops, the colour Doppler signals formed a complete ring of colour in the bowel wall (Fig. 9.19a). In others, the flow in the supplying subserosal mesenteric vessel and bowel wall had a Y-shaped appearance (Fig. 9.19b).