- Email: [email protected]
Ultrasonic doppler studies of hepatocellular carcinoma and comparison with other hepatic focal lesions
GASTROENTRROLOGY
1989;97:1489-97
Ultrasonic Doppler Studies of Hepatocellular Carcinoma and Comparison With Other Hepatic Focal Lesions KUNIHIKO OHNISHI and FUMIO NOMURA First Department
of Medicine,
Chiba University
One hundred fifty-four liver lesions, including 63 hepatocellular carcinomas, were studied to determine the value of duplex ultrasound in the diagnosis of small hepatocellular carcinomas. Arterial Doppler signals were obtained either within the body of the tumor, at its periphery, or in both locations, from 26 of 37 hepatocellular carcinomas 53 cm in diameter and from all 26 hepatocellular carcinomas with a diameter >3 cm. Arterial Doppler signals were obtained at the periphery of 5 of 7 4 of 11 liver metastatic tucholangiocarcinomas, mors, and 5 of 23 hemangiomas. No such signals were obtained from 29 regenerative nodules, 10 hepatic pseudotumors, and 11 liver cysts. The mean peak systolic frequency seen in hepatocellular carcinoma (1.2 kHz) was significantly greater than in cholangiocarcinoma (0.6 kHz), metastatic tumors (0.5 kHz), or hemangiomas (0.3 kHz). A peak systolic frequency of >3 kHz was found in 6 of 8 hepatocellular carcinomas 24 cm in diameter with angiographically proven arterioportal shunting, whereas the value in other hepatocellular carcinomas or other hepatic focal lesions was C2.6 kHz. This study showed that the peak systolic shift was related to the degree of arterioportal shunting. Because shunting is either minor or nonexistent in small hepatocellular carcinomas, the value of duplex Doppler ultrasound in the diagnosis of these lesions appears to be limited.
R
ecent progress in ultrasound (US) has provided the means to detect small hepatocellular carcinomas (l,Z), which has helped to improve the prognosis of this condition by allowing the application of various newly developed and improved therapeutic modalities (3~). However, the separation of small carcinomas 53 cm in diameter from other hepatic focal lesions is often difficult even in the hands of an expert. It is markedly more so in the presence of
School of Medicine,
Chiba, Japan
diffuse hepatocellular disease and liver cirrhosis. The differential diagnosis includes focal fatty infiltration (5,6), pseudotumors (7,8), cholangiocarcinoma, liver metastases, hemangiomas (9), and regenerative nodules (adenomatous hyperplasia) (10-15). Recently, Taylor and Burns (16)reported that Doppler signals could be detected from abdominal malignant tumors using US and pulsed Doppler instruments operating at a center frequency of 3.0 MHz. Subsequently, Taylor et al. (17) reported the results of Doppler examination of 68 liver lesions, including 12 hepatocellular carcinomas. Angiographic findings suggested that high velocity Doppler signals might be associated with arterioportal shunting. However, it remained to be proven whether duplex Doppler US could detect small carcinomas or not, because the smallest tumor in Taylor’s series was 4 cm in diameter (17), and it is known that arterioportal shunting is rarely found in hepatocellular carcinomas of 13 cm in diameter (i8,19). Therefore, as an extension of their work, we investigated whether duplex Doppler US was useful in the diagnosis of small hepatocellular carcinomas 13 cm in diameter as well as for larger lesions. Subjects
and Methods
Subjects Our subjects consisted of 94 men and 28 women who ranged in age from 38 to 80 yr. Forty-eight subjects had hepatocellular carcinoma: 10 had two tumors, 1 had three tumors, 1 had four tumors, and all the others had one tumor each. Two patients had cholangiocarcinoma: one had three tumors and the other had four tumors. The diagnosis of hepatocellular carcinoma or cholangiocarci-
Abbreviation
used in this paper: US, ultrasound.
0 1989 by the American
Gastroenterological
QQ16-5Q05/89/$3.5Q
Association
1490
GASTROENTEROLOGY
OHNISHI AND NOMURA
noma was made histologically using surgically resected specimens in 3 subjects and specimens obtained by USguided percutaneous liver biopsy with 1% or 21-gauge needles in the other subjects. Tumor size was assessed from the hepatic angiogram in 34 of the 48 subjects with hepatocellular carcinoma, and by the use of US or computed tomography in the other subjects with hepatocellular carcinoma or cholangiocarcinoma. Five subjects had gastric carcinoma with liver metastasis: 2 had one metastasis, 2 had two metastases, and 1 had three metastases. One subject had gallbladder carcinoma with two liver metastases. The diagnosis of gastric carcinoma, gallbladder carcinoma, or liver metastasis was made histologically using specimens obtained from the tumors. The size of the liver metastases was assessed from the B-image on US. Twenty-one subjects had hemangiomas: 2 patients had two lesions and the others had one each. Hemangioma was diagnosed by hepatic angiogram in 8 subjects and by dynamic computed tomography, using intravenous bolus injection of an iodinated contrast agent (20), in 13 subjects. Assessment of hemangioma size was made either from the angiogram or with dynamic computed tomography. Ten subjects had hepatic pseudotumors on US, which showed as a solitary hypoechoic area within the liver parenchyma with diffuse regions of increased echogenicity. In these 10 subjects US-guided percutaneous liver biopsy of the hypoechoic area was undertaken with 1%gauge needles and biopsy specimens of the surrounding liver were obtained using 16- or 17-gauge needles. Microscopic examination of specimens from the hypoechoic regions revealed normal hepatic parenchymal cells, whereas tissue samples from the surrounding livers showed high fat levels. Pseudotumor size was assessed using the B-image on US. Twentyfour subjects had regenerative nodules (IO-15), which were depicted as focal hypoechoic areas with a relatively distinct margin on US: 5 subjects had two regenerative nodules and the others had one regenerative nodule each. Ultrasound-guided percutaneous liver biopsy was undertaken using 1% or 21-gauge needles for the hypoechoic areas and 16- or ‘17-gauge needles for the surrounding liver in these 24 subjects. Microscopy of specimens from the hypoechoic regions revealed normal hepatocytes and no malignant cells, whereas samples from the other regions showed liver cirrhosis. Hepatic angiography was also performed in 16 of these subjects but the hypoechoic areas shown on US could not be distinguished from the surrounding liver by angiography. Follow-up US examination was performed in these 24 subjects with regenerative nodule(s) from 6 mo to 6 yr after the initial sonograms were obtained. No change has been demonstrated in the size of the regenerative nodules in this period of time. Eleven subjects had liver cysts, the size of which was assessed by their B-image on US.
Ultrasonic
Technique
and
Examination
All US and Doppler flow examinations were performed with a duplex system that consisted of a 3.0-MHz electronic real-time convex scanner and a pulsed Doppler device operating at a center frequency of 3.0 MHz and a
Vol. 97, No. 6
pulse repetition frequency of 4.2 kHz (Aloka SSD-650 and modified UGR-650; Aloka Co., Tokyo, Japan). Informed written consent was obtained from each patient. The maximum sample volume (10 mm) was used to search for tumor signals. The Doppler filter was set at 100 Hz and it was able to be changed manually. The vessels or hepatic lesions in question were first identified by a real-time scanner, and the line of sight of the transducer along with the Doppler range gate (sample volume) was superimposed on the B-scan, which was frozen. The sample volume cursor was moved to detect Doppler shift signals from within, at the periphery of, or around the hepatic lesion or from within the vessel, while the operator listened to the audio output of the Doppler-shifted signal. Because the imaging and the Doppler examination were not simultaneous, the image was updated at regular intervals to ensure that the sample volume remained within the abnormal region. When a Doppler signal was detected from the tumor, the sample volume was reduced to 2 mm to determine more precisely the location of the signal.
Angiographic
Examination
Celiac, selective hepatic, and superior mesenteric angiography was performed in 44 subjects with hepatocellular carcinoma, 2 with cholangiocarcinoma, 8 with hemangioma, and 16 with regenerative nodules several days after the US and Doppler examination. The methods used have been described (21). Vascularity of the hepatocellular carcinomas was graded as follows: 0 = avascular, 1 = hypovascular, 2 = isovascular, 3 = vascular, and 4 = highly vascular.
Statistical
Analysis
Results are given as the mean and the range. Comparisons were made by the Mann-Whitney U-test and the 2 test with Yates’ correction, and a p value ~0.05 was considered significant.
Results Arterial Doppler signals were obtained from 54 of 63 hepatocellular carcinomas (mean diameter, 3.1 cm; range, 1.0-8.7 cm). They were obtained from within the body of 37 lesions (Figures l-4) and at the periphery of 47 lesions (Table 1). Arterial Doppler signals could be obtained only at the periphery of 5 of 7 cholangiocarcinomas (mean diameter, 1.7 cm; range, 1.6-3.0 cm), in 4 of 11 metastatic tumors (mean diameter, 3.2 cm; range, 1.6-6.0 cm), and in 5 of 23 hemangiomas (mean diameter, 2.3 cm; range, 1.0-4.4 cm). They were not obtained at all from regenerative nodules (mean diameter, 1.5 cm; range, 0.7-2.3 cm) (Figure 5), pseudotumors (mean diameter, 2.4 cm; range, 1.4-3.2 cm) (Figure 6), and liver cysts (mean diameter, 2.6 cm; range, 1.0-6.5 cm). No Doppler signals were detected from the hepatic pa-
December 1989
Figure
DOPPLER STUDIES OF HEPATOCELLULAR CARCINOMA
1491
1. Hepatocellular carcinoma 1.5 cm in diameter. A. A hepatocellular carcinoma in the lateral subsegment of the left lobe is visualized as a hypoechoic area (arrow] by US B-image. The white bar shows the position of the 2-mm sample volume. B. Arterial and portal Doppler signals within the tumor.
renchyma other than from the focal lesions. Doppler signals were obtained, however, from the major and segmental branches of the portal vein, from the branches of the hepatic artery [which normally run with the branches of the portal vein and are rarely visualized by B-image on US), and from the right, middle, and left hepatic veins. The mean peak systolic frequency for hepatocellular carcinoma was 1.2 kHz (range, O-4.2 kHz), which was significantly greater than those for cholangiocarcinoma (0.6 kHz; range, o-l.5 kHz, p < 0.05),liver metastases (0.5 kHz; range, O-2.2 kHz, p < 0.05), and hemangiomas (0.3kHz; range, o-l.7 kHz, p < 0.05). Arterial Doppler signals were obtained from within the body, at the periphery, or from both sites in 28 of 37 small hepatocellular carcinomas 13 cm in diameter and from all 26 carcinomas with a diameter >3 cm. They were obtained within the body of 17 of 37 small carcinomas and 20 of 26 larger carcinomas, and at the periphery of the tumor in 26 small and 21 larger lesions, respectively. The mean peak systolic frequency of the small carcinomas was significantly less than that of the larger carcinomas, and the
vascularity was also significantly less in the smaller tumors. Arterioportal shunting was observed angiographically in 1 of 34 small hepatocellular carcinomas studied, whereas it was detected in 8 of 25 larger carcinomas (Table 1). Arterial Doppler signals were more frequently obtained from tumors with greater vascularity, with the isovascular tumors showing the lowest number of Doppler signals (Table 2). The mean peak systolic frequency increased in the order of isovascular hepatocellular carcinoma, vascular carcinoma, and highly vascular carcinoma. A peak systolic frequency of 23 kHz was seen in 6 of 8 hepatocellular carcinomas r4 cm in diameter with arterioportal shunting, whereas the peak systolic frequency was C2.6 kHz in the other hepatocellular carcinomas and in other hepatic focal lesions. The mean diameter of hepatocellular carcinomas with arterioportal shunting was 5.4 cm (range, 1.3-8.7 cm), and arterial Doppler signals were observed both from within and at the periphery of these tumors. The mean peak systolic frequency of the carcinomas with shunting was 3.0 kHz (range, 1.6-4.2kHz).
1492
OHNISHI AND NOMURA
GASTROENTEROLOGY Vol. 97, No. 6
Figure 2. Hepatocellular carcinoma 2.0 cm in diameter. A. A hepatocellular carcinoma in the posterior inferior subsegment of the right lobe is visualized as a hypoechoic area by US-B image (arrow). The white bar shows the position of the lo-mm sample volume. B. Arterial Doppler signal of 1.4 kHz from within the tumor. C. Proper hepatic arteriogram showing a vascular mass (arrow] in the right lobe of the liver.
December 1989
DOPPLER STUDIES OF HEPATOCELLULAR CARCINOMA
1493
Figure 3. Hepatocellular carcinoma 4.0 cm in diameter. A. A hepatocellular carcinoma in the posterior inferior subsegment of the right lobe is visualized as a hyperechoic area by US B-image (arrow]. The white bar shows the position of the lo-mm sample volume. B. Arterial Doppler signal of 1.5 kHz from within the tumor. C. Right hepatic arteriogram showing a vascular mass (arrow) in the right lobe of the liver.
1494
OHNISHI AND NOMURA
GASTROENTEROLOGY Vol. 97, No. 6
Figure 4. Hepatocellular carcinoma 7 cm in diameter. A. A hepatocellular carcinoma in the medial subsegment of the left lobe is visualized as a heterogeneous area by US Bimage. The white bar shows the position of the 2-mm sample volume. B. Arterial Doppler signal of 4.2 kHz from within the tumor. C. Proper hepatic arteriogram showing a highly vascular mass (arrow).
December 1989
DOPPLERSTUDIESOF HEPATOCELLULAR CARCINOMA 1495
Table 1. Ultrasound, Doppler, and Angiographic Characteristics of Hepatocellular Carcinomas Hepatocellular Group Mean diameter (range] (cm) B-image of hepatocellular carcinoma Hypoechoic Isoechoic with low echoic periphery Hyperechoic Heterogeneous Arterial Doppler signals obtained Within the tumor At the periphery of the tumor Within or at the periphery of the tumor or both sites Mean peak systolic frequency of arterial Doppler signal (range) (kHz1 Mean vascularity (range) Arterioportal
shunting
53 cm (n = 37) 1.9 (1.0-3.0)
carcinomas >3 cm (n = 26) 4.9 (3.2-8.7)
27 (73%) 4 (11%)
5 (19%) 0 (0%)
3 (8%) 3 (8%)
6 (23%) 15 (58%)
17(46%)
26 (70%)
20(77%) 21(81%)
28 (76%)
26 (100%)
0.7 (O-1.6)a
1.9 (0.4-4.2)a
2.7 (2-3)” n = 34 1 (3%)b
3.4 (2-4)O n = 25 8 (32%)b
Angiography was performed in 44 patients with 59 hepatocellular carcinomas. The grading of vascularity was as follows: 0 = avascular, 1 = hypovascular, 2 = isovascular, 3 = vascular, 4 = highly vascular. ’ p < 0.05. b p < 0.01.
Discussion In 1977, Wells et al. (22) reported that from among 9 patients with palpable masses in their breasts they elicited Doppler-shifted signals with a systolic modulation from 3 patients with malignancies, using an ~-MHZ continuous-wave Doppler system. They found no signals or found weak signals in the remaining 6 patients who had benign lesions. They suggested that this technique could distinguish benign breast lesions from malignant ones and that it Table 2. Ultrasound and Doppler Characteristics
might be possible to develop a new high-speed ultrasonic Doppler scanner to make breast scanning practicable. Since then breast cancer diagnosis by Doppler US has been studied by many researchers (23-27). White and Cledgett (23), Lypacewicz et al. (25), and Burns et al. (27) observed Doppler signals with systolic modulation from benign as well as malignant breast lesions. Gros et al. (24) reported that they obtained Doppler signals from every “inflammatory and evolutive lesion benign or malignant.” Regarding the spectra of the signals, in breast cancer, when lo-MHz continuous-wave Doppler units were used in particular, -90% of localized tumors exhibited abnormal properties (27). Among these were very high peak velocities, increased mean velocities, and forward and reverse flow (23-27). From these observations, it would seem that it is the quality of the Doppler signal rather than its presence or absence that differentiates malignant breast tumors from benign lesions (23,27). In 1979, Mountford and Atkinson (28) tested this technique in abdominal tumors, but found they could not obtain Doppler signals from abdominal tumors using a lo-MHz continuous-wave Doppler device. They concluded that this was because the high frequency of the ultrasound limited the penetration to 30 mm because most of the energy was attenuated in the abdominal wall. Recent technological developments in the duplex US and pulsed Doppler systems have made possible the noninvasive measurement of flow velocity and blood flow in deep abdominal and pelvic vessels (16). In 1985, Taylor and Burns (16) published the first report on obtaining abnormal signals using pulsed Doppler instruments operating at a center frequency of 3.0 MHz from several different malignant abdominal tumors, including hepatoblastoma, metastatic neuroblastoma, a mixed Mullerian cell tumor of the uterus, and a renal carcinoma. Subsequently, in 1987, Taylor et al. (17) reported the
of Isovascular,
Vascular,
and Highly Vascular
Hepatocellular
Carcinomas Hepatocellular
Group Mean diameter (range) (cm) Arterial Doppler signals obtained Within the tumor At the periphery of the tumor Within or at the periphery of tumor or both sites Mean peak systolic frequency of arterial Doppler signal (range) (kHz) Arterioportal shunting
Isovascular (n = 12)
carcinomas
Vascular (n = 35)
Highly vascular (n = 12)
2.0 (1.0-3.6)
2.7 (1.0-6.0)
5.7 (4.0-8.7)a.b
4 (33%) 5 (42%) 6 (50%)
17 (49%) 27 (77%) 32 (91%)
11(92%) 12 (100%)
0.5 (O-1.6)
1.0 (O-3.1)
2.5 (0.7-l.2)",b
0 (0%)
2 (6%)
11(92%)
7 (58%)
Angiography was performed in 44 patients with 59 hepatocellular carcinomas. The grading of vascularity was as follows: 0 = avascular, 1 = hypovascular, 2 = isovascular, 3 = vascular, 4 = highly vascular. 0 p < 0.05 vs. isovascular group. b p < 0.05 vs. vascular group.
1496
OHNISHI AND NOMURA
GASTROENTEROLOGY Vol. 97, No. 6
periphery of hepatocellular carcinomas, other hepatic focal lesions, and hepatic areas other than focal lesions? (b) Are the Doppler signals associated with hepatocellular carcinoma different from those from the other hepatic focal lesions or those from the other parts of the liver? The present study demonstrated that arterial Doppler signals were obtained either from within the body, at the periphery, or from both regions in about 75% of small hepatocellular carcinomas and in all the larger carcinomas, Arterial Doppler signals were obtained only at the periphery of some cholangiocarcinemas, some liver metastatic tumors, and some hemangiomas, whereas they were not obtained at all from regenerative nodules, pseudotumors, and liver cysts. This result seemed to be in accord with the angiographic findings of these hepatic focal lesions. Most hepatocellular carcinomas are vascular both within the body and at the periphery. Hemangiomas, cholangiocarcinomas, and metastases are more vascular at the periphery. Regenerative nodules and pseudotumors are isovascular, whereas liver cysts
Figure 5. Regenerative nodule 2.0 cm in diameter. A regenerative nodule in the medial subsegment of the left lobe is visualized as a hypoechoic area by US B-image (arrows]. The white bar shows the position of the z-mm sample volume.
results of Doppler examination in 68 consecutive patients with focal liver lesions, including 12 hepatocellular carcinomas that were all >4 cm in diameter. All 10 tumors with Doppler shifts of ~5 kHz proved to be hepatocellular carcinomas, and the correlation with angiographic findings suggested that the high-velocity Doppler signals were associated with large pressure gradients due to arterioportal shunting. These reports encouraged us to determine whether duplex Doppler US was also useful for the diagnosis of small hepatocellular carcinomas 53 cm in diameter as well as for tumors with a diameter >3 cm. Taylor et al. did not state whether Doppler signals were obtained from within the tumors or at their peripheries because most of the carcinomas they studied were diffuse or multifocal. However, we believed that it might be important for the diagnosis of small hepatocellular carcinomas to determine the precise origin of the Doppler signal. Therefore, we investigated two main areas. (a) Can a pulsed Doppler device operating at a center frequency of 3.0 MHz detect Doppler signals from within or at the
Figure 6. Pseudotumor 1.8 cm in diameter. A pseudotumor is ... . me .. liver .. ’ ’ area wifnin visualized as a h!qpoecnoic parenchyma with diffl use increased echogenicity by US Ble White image (arrow). Tl._ .._____bar shows the position of the Z-mm sample volume.
December 1989
are avascular. No Doppler signals were obtained from the hepatic parenchyma other than from the focal lesions, which differed somewhat from the findings of Taylor et al. that Doppler signals of 50.7 kHz were obtainable from the hepatic parenchyma (I 7). We found that the mean peak systolic frequency for hepatocellular carcinoma was significantly greater than for cholangiocarcinoma and hemangioma, which was in accordance with the report of Taylor et al. (17). The mean peak systolic frequency of the hepatocellular carcinomas was related to their size, vascularity, and degree of arterioportal shunting, with the peak frequency being significantly less in smaller tumors. A peak systolic frequency of 23 kHz was obtained only in hepatocellular carcinomas 24 cm in diameter with arterioportal shunting and the peak frequency was C2.6 kHz in smaller carcinomas and in other hepatic focal lesions. These findings led us to conclude that duplex Doppler US is of limited value in the diagnosis of hepatic focal lesions including small hepatocellular carcinomas. When arterial Doppler signals are obtained from within the body of a hepatic focal lesion it is probably a hepatocellular carcinoma, and when the signals are obtained at the periphery it is likely to be a tumor. However, the absence of arterial Doppler signals does not necessarily mean that the lesion is not a tumor. Measurement of the peak systolic frequency is thus of little value in the differentiation of small hepatocellular carcinomas from other lesions encountered clinically.
DOPPLER STUDIES OF HEPATOCELLULAR CARCINOMA
11.
12.
13.
14.
15. 16. 17.
18.
19.
20.
21.
22.
References 1.
Sheu JC, Sung JL, Chen DS, et al. Ultrasonography of small hepatic tumors using high resolution linear array real-time instruments. Radiology 1984;150:797-802. 2. Shinagawa T, Ohto M, Kimura K, et al. Diagnostic and clinical features of small hepatocellular carcinoma with emphasis on the utility of real-time ultrasonography: a study of 51 patients. Gastroenterology 1984;86:496-502. 3. Okuda K, Ohtsuki T, Obata H, et al. Natural history of hepatocellular carcinoma and prognosis in relation to treatment. Study in 850 patients. Cancer 1985;56:918-28. 4. Ohnishi K, Tanabe Y, Ryu M, et al. Prognosis of hepatocellular carcinoma smaller than 5 cm in relation to treatment: study of 100 patients. Hepatology 1987;7:1285-90. 5. Yates CK, Streight R. Focal fatty infiltration of the liver simulating metastatic diseases. Radiology 1986;159:83-4. 6. Kawashima A, Suehiro S, Murayama S, Russell WJ. Focal fatty infiltration of the liver mimicking a tumor: sonographic and CT features. J Comput Assisted Tomogr 1986;10:329-31. 7. Berland LL. Focal areas of decreased echogenicity in the liver at the porta hepatis. J Ultrasound Med 1986:5:157-g. 8. White EM, Simeone JF, Mueller PR, Grant EG, Choyke PL, Zeman RK. Focal periportal sparing in hepatic fatty infiltration: a cause of hepatic pseudomass on US. Radiology 1987;162:57-9. 9. Bree RL, Schwab RE, Neiman HL. Solitary echogenic spot in the liver: is it diagnostic of a hemangioma? Am J Roentgen01 1983;140:41-5. HA. Benign epithelial tumors and tumorlike 10. Edmondson
23. 24.
1497
lesions of the liver. In: Okuda K, Peters RL, eds. Hepatocellular carcinoma. New York: Wiley, 1976:309-30. Sasaki K. Adenomatous hyperplasia in liver cirrhosis: an approach from a microangiographical point of view. Gann Monogr Cancer Res 1980;25:127-40. Makuuchi M, Hasegawa H, Yamazaki S. Development of liver surgery by operative sonography. Semin Gastroenterol Surg 1983;13:151-73 (in Japanese). Akimoto S, Saito A, Watayoh T, Isobe Y, Takasaki T. Diagnostic capability and limitation of ultrasonography in small hepatocellular carcinoma. Diagn Imaging Abdomen 1984; 4:209-16 (in Japanese]. Arakawa M, Kage M, Sugihara S, Nakashima T, Suenaga M, Okuda K. Emergence of malignant lesions within an adenomatous hyperplastic nodule in a cirrhotic liver. Observations in five cases. Gastroenterology 1986;91:198-208. Itai Y, Ohnishi S, Ohtomo K, et al. Regenerative nodules of liver cirrhosis: MR imaging. Radiology 1987;165:419-23. Taylor KJW, Burns PN. Duplex Doppler scanning in the pelvis and abdomen. Ultrasound Med Biol 1985;11:643-58. Taylor KJW, Ramos I, Morse SS, Fortune KL, Hammers L, Taylor CR. Focal liver masses: differential diagnosis with pulsed Doppler US. Radiology 1987;164:643-7. Yamasaki S, Hasegawa H, Makuuchi M. Clinicopathological observation of minute liver cancer and a new method of hepatectomy: analysis of 27 resected cases. Acta Hepatol Jpn 1981;22:1714-24 (in Japanese). Takayasu K, Shima Y, Muramatsu Y, et al. Angiography of small hepatocellular carcinoma: analysis of 105 resected tumors. Am J Roentgen01 1986;147:525-9. Nakayama T, Hiyama Y, Ohnishi K, et al. Arterioportal shunts on dynamic computed tomography. Am J Roentgen01 1983; 140:953-7. Ohnishi K, Okuda K, Ohtsuki T, et al. Formation of hilar collaterals or cavernous transformation after portal vein obstruction by hepato-cellular carcinoma. Observations in ten patients. Gastroenterology 1984;87:1156-3. Wells PNT, Halliwell M, Skidmore R, Webb AJ, Woodcock JP. Tumor detection by ultrasonic Doppler blood-flow signals. Ultrasonics 1977;15:231-2. White DN, Cledgett PR. Breast carcinoma detection by ultrasonic Doppler signals. Ultrasound Med Biol 1978;4:329-35. Gros Ch M, Dale G, Gairand B. Breast echography: criteria of malignancy and results. In: Kurjak A, ed. Recent advances in ultrasound diagnosis. Amsterdam: Excerpta Medica, 1978:
292-8. 25. Lypacewicz
G, Powalowski T, Lukawska K. Ultrasonic examination of breast tumors with Doppler method. In: Fillipczynski L, Zieniuk JK, eds. Proceedings of the Second Congress of the Federation of Acoustical Societies of Europe. Volume II. Polish Academy of Sciences, 1978:153-6. assessment of an 26. Minasian H, Bamber JC. A preliminary ultrasonic Doppler method for the study of blood flow in human breast cancer. Ultrasound Med Biol 1982;8:357-64. 27. Burns PN, Halliwell M, Wells PNT. Ultrasonic Doppler studies of the breast. Ultrasound Med Biol 1982;8:127-43. 28. Mountford RA, Atkinson P. Doppler ultrasound examination of pathologically enlarged lymph nodes. Br J Radio1 1979; 52:464-7.
Received February 17, 1988. Accepted June 2, 1989. Address requests for reprints to: Kunihiko Ohnishi, M.D., First Department of Medicine, Chiba University School of Medicine, 1-8-1 Inohana-cho, Chiba City, Chiba 280, Japan. This study was supported in part by a research grant from the Intractable Diseases Division of the Public Health Bureau, Ministry of Health and Welfare, Japan.
1989;97:1489-97
Ultrasonic Doppler Studies of Hepatocellular Carcinoma and Comparison With Other Hepatic Focal Lesions KUNIHIKO OHNISHI and FUMIO NOMURA First Department
of Medicine,
Chiba University
One hundred fifty-four liver lesions, including 63 hepatocellular carcinomas, were studied to determine the value of duplex ultrasound in the diagnosis of small hepatocellular carcinomas. Arterial Doppler signals were obtained either within the body of the tumor, at its periphery, or in both locations, from 26 of 37 hepatocellular carcinomas 53 cm in diameter and from all 26 hepatocellular carcinomas with a diameter >3 cm. Arterial Doppler signals were obtained at the periphery of 5 of 7 4 of 11 liver metastatic tucholangiocarcinomas, mors, and 5 of 23 hemangiomas. No such signals were obtained from 29 regenerative nodules, 10 hepatic pseudotumors, and 11 liver cysts. The mean peak systolic frequency seen in hepatocellular carcinoma (1.2 kHz) was significantly greater than in cholangiocarcinoma (0.6 kHz), metastatic tumors (0.5 kHz), or hemangiomas (0.3 kHz). A peak systolic frequency of >3 kHz was found in 6 of 8 hepatocellular carcinomas 24 cm in diameter with angiographically proven arterioportal shunting, whereas the value in other hepatocellular carcinomas or other hepatic focal lesions was C2.6 kHz. This study showed that the peak systolic shift was related to the degree of arterioportal shunting. Because shunting is either minor or nonexistent in small hepatocellular carcinomas, the value of duplex Doppler ultrasound in the diagnosis of these lesions appears to be limited.
R
ecent progress in ultrasound (US) has provided the means to detect small hepatocellular carcinomas (l,Z), which has helped to improve the prognosis of this condition by allowing the application of various newly developed and improved therapeutic modalities (3~). However, the separation of small carcinomas 53 cm in diameter from other hepatic focal lesions is often difficult even in the hands of an expert. It is markedly more so in the presence of
School of Medicine,
Chiba, Japan
diffuse hepatocellular disease and liver cirrhosis. The differential diagnosis includes focal fatty infiltration (5,6), pseudotumors (7,8), cholangiocarcinoma, liver metastases, hemangiomas (9), and regenerative nodules (adenomatous hyperplasia) (10-15). Recently, Taylor and Burns (16)reported that Doppler signals could be detected from abdominal malignant tumors using US and pulsed Doppler instruments operating at a center frequency of 3.0 MHz. Subsequently, Taylor et al. (17) reported the results of Doppler examination of 68 liver lesions, including 12 hepatocellular carcinomas. Angiographic findings suggested that high velocity Doppler signals might be associated with arterioportal shunting. However, it remained to be proven whether duplex Doppler US could detect small carcinomas or not, because the smallest tumor in Taylor’s series was 4 cm in diameter (17), and it is known that arterioportal shunting is rarely found in hepatocellular carcinomas of 13 cm in diameter (i8,19). Therefore, as an extension of their work, we investigated whether duplex Doppler US was useful in the diagnosis of small hepatocellular carcinomas 13 cm in diameter as well as for larger lesions. Subjects
and Methods
Subjects Our subjects consisted of 94 men and 28 women who ranged in age from 38 to 80 yr. Forty-eight subjects had hepatocellular carcinoma: 10 had two tumors, 1 had three tumors, 1 had four tumors, and all the others had one tumor each. Two patients had cholangiocarcinoma: one had three tumors and the other had four tumors. The diagnosis of hepatocellular carcinoma or cholangiocarci-
Abbreviation
used in this paper: US, ultrasound.
0 1989 by the American
Gastroenterological
QQ16-5Q05/89/$3.5Q
Association
1490
GASTROENTEROLOGY
OHNISHI AND NOMURA
noma was made histologically using surgically resected specimens in 3 subjects and specimens obtained by USguided percutaneous liver biopsy with 1% or 21-gauge needles in the other subjects. Tumor size was assessed from the hepatic angiogram in 34 of the 48 subjects with hepatocellular carcinoma, and by the use of US or computed tomography in the other subjects with hepatocellular carcinoma or cholangiocarcinoma. Five subjects had gastric carcinoma with liver metastasis: 2 had one metastasis, 2 had two metastases, and 1 had three metastases. One subject had gallbladder carcinoma with two liver metastases. The diagnosis of gastric carcinoma, gallbladder carcinoma, or liver metastasis was made histologically using specimens obtained from the tumors. The size of the liver metastases was assessed from the B-image on US. Twenty-one subjects had hemangiomas: 2 patients had two lesions and the others had one each. Hemangioma was diagnosed by hepatic angiogram in 8 subjects and by dynamic computed tomography, using intravenous bolus injection of an iodinated contrast agent (20), in 13 subjects. Assessment of hemangioma size was made either from the angiogram or with dynamic computed tomography. Ten subjects had hepatic pseudotumors on US, which showed as a solitary hypoechoic area within the liver parenchyma with diffuse regions of increased echogenicity. In these 10 subjects US-guided percutaneous liver biopsy of the hypoechoic area was undertaken with 1%gauge needles and biopsy specimens of the surrounding liver were obtained using 16- or 17-gauge needles. Microscopic examination of specimens from the hypoechoic regions revealed normal hepatic parenchymal cells, whereas tissue samples from the surrounding livers showed high fat levels. Pseudotumor size was assessed using the B-image on US. Twentyfour subjects had regenerative nodules (IO-15), which were depicted as focal hypoechoic areas with a relatively distinct margin on US: 5 subjects had two regenerative nodules and the others had one regenerative nodule each. Ultrasound-guided percutaneous liver biopsy was undertaken using 1% or 21-gauge needles for the hypoechoic areas and 16- or ‘17-gauge needles for the surrounding liver in these 24 subjects. Microscopy of specimens from the hypoechoic regions revealed normal hepatocytes and no malignant cells, whereas samples from the other regions showed liver cirrhosis. Hepatic angiography was also performed in 16 of these subjects but the hypoechoic areas shown on US could not be distinguished from the surrounding liver by angiography. Follow-up US examination was performed in these 24 subjects with regenerative nodule(s) from 6 mo to 6 yr after the initial sonograms were obtained. No change has been demonstrated in the size of the regenerative nodules in this period of time. Eleven subjects had liver cysts, the size of which was assessed by their B-image on US.
Ultrasonic
Technique
and
Examination
All US and Doppler flow examinations were performed with a duplex system that consisted of a 3.0-MHz electronic real-time convex scanner and a pulsed Doppler device operating at a center frequency of 3.0 MHz and a
Vol. 97, No. 6
pulse repetition frequency of 4.2 kHz (Aloka SSD-650 and modified UGR-650; Aloka Co., Tokyo, Japan). Informed written consent was obtained from each patient. The maximum sample volume (10 mm) was used to search for tumor signals. The Doppler filter was set at 100 Hz and it was able to be changed manually. The vessels or hepatic lesions in question were first identified by a real-time scanner, and the line of sight of the transducer along with the Doppler range gate (sample volume) was superimposed on the B-scan, which was frozen. The sample volume cursor was moved to detect Doppler shift signals from within, at the periphery of, or around the hepatic lesion or from within the vessel, while the operator listened to the audio output of the Doppler-shifted signal. Because the imaging and the Doppler examination were not simultaneous, the image was updated at regular intervals to ensure that the sample volume remained within the abnormal region. When a Doppler signal was detected from the tumor, the sample volume was reduced to 2 mm to determine more precisely the location of the signal.
Angiographic
Examination
Celiac, selective hepatic, and superior mesenteric angiography was performed in 44 subjects with hepatocellular carcinoma, 2 with cholangiocarcinoma, 8 with hemangioma, and 16 with regenerative nodules several days after the US and Doppler examination. The methods used have been described (21). Vascularity of the hepatocellular carcinomas was graded as follows: 0 = avascular, 1 = hypovascular, 2 = isovascular, 3 = vascular, and 4 = highly vascular.
Statistical
Analysis
Results are given as the mean and the range. Comparisons were made by the Mann-Whitney U-test and the 2 test with Yates’ correction, and a p value ~0.05 was considered significant.
Results Arterial Doppler signals were obtained from 54 of 63 hepatocellular carcinomas (mean diameter, 3.1 cm; range, 1.0-8.7 cm). They were obtained from within the body of 37 lesions (Figures l-4) and at the periphery of 47 lesions (Table 1). Arterial Doppler signals could be obtained only at the periphery of 5 of 7 cholangiocarcinomas (mean diameter, 1.7 cm; range, 1.6-3.0 cm), in 4 of 11 metastatic tumors (mean diameter, 3.2 cm; range, 1.6-6.0 cm), and in 5 of 23 hemangiomas (mean diameter, 2.3 cm; range, 1.0-4.4 cm). They were not obtained at all from regenerative nodules (mean diameter, 1.5 cm; range, 0.7-2.3 cm) (Figure 5), pseudotumors (mean diameter, 2.4 cm; range, 1.4-3.2 cm) (Figure 6), and liver cysts (mean diameter, 2.6 cm; range, 1.0-6.5 cm). No Doppler signals were detected from the hepatic pa-
December 1989
Figure
DOPPLER STUDIES OF HEPATOCELLULAR CARCINOMA
1491
1. Hepatocellular carcinoma 1.5 cm in diameter. A. A hepatocellular carcinoma in the lateral subsegment of the left lobe is visualized as a hypoechoic area (arrow] by US B-image. The white bar shows the position of the 2-mm sample volume. B. Arterial and portal Doppler signals within the tumor.
renchyma other than from the focal lesions. Doppler signals were obtained, however, from the major and segmental branches of the portal vein, from the branches of the hepatic artery [which normally run with the branches of the portal vein and are rarely visualized by B-image on US), and from the right, middle, and left hepatic veins. The mean peak systolic frequency for hepatocellular carcinoma was 1.2 kHz (range, O-4.2 kHz), which was significantly greater than those for cholangiocarcinoma (0.6 kHz; range, o-l.5 kHz, p < 0.05),liver metastases (0.5 kHz; range, O-2.2 kHz, p < 0.05), and hemangiomas (0.3kHz; range, o-l.7 kHz, p < 0.05). Arterial Doppler signals were obtained from within the body, at the periphery, or from both sites in 28 of 37 small hepatocellular carcinomas 13 cm in diameter and from all 26 carcinomas with a diameter >3 cm. They were obtained within the body of 17 of 37 small carcinomas and 20 of 26 larger carcinomas, and at the periphery of the tumor in 26 small and 21 larger lesions, respectively. The mean peak systolic frequency of the small carcinomas was significantly less than that of the larger carcinomas, and the
vascularity was also significantly less in the smaller tumors. Arterioportal shunting was observed angiographically in 1 of 34 small hepatocellular carcinomas studied, whereas it was detected in 8 of 25 larger carcinomas (Table 1). Arterial Doppler signals were more frequently obtained from tumors with greater vascularity, with the isovascular tumors showing the lowest number of Doppler signals (Table 2). The mean peak systolic frequency increased in the order of isovascular hepatocellular carcinoma, vascular carcinoma, and highly vascular carcinoma. A peak systolic frequency of 23 kHz was seen in 6 of 8 hepatocellular carcinomas r4 cm in diameter with arterioportal shunting, whereas the peak systolic frequency was C2.6 kHz in the other hepatocellular carcinomas and in other hepatic focal lesions. The mean diameter of hepatocellular carcinomas with arterioportal shunting was 5.4 cm (range, 1.3-8.7 cm), and arterial Doppler signals were observed both from within and at the periphery of these tumors. The mean peak systolic frequency of the carcinomas with shunting was 3.0 kHz (range, 1.6-4.2kHz).
1492
OHNISHI AND NOMURA
GASTROENTEROLOGY Vol. 97, No. 6
Figure 2. Hepatocellular carcinoma 2.0 cm in diameter. A. A hepatocellular carcinoma in the posterior inferior subsegment of the right lobe is visualized as a hypoechoic area by US-B image (arrow). The white bar shows the position of the lo-mm sample volume. B. Arterial Doppler signal of 1.4 kHz from within the tumor. C. Proper hepatic arteriogram showing a vascular mass (arrow] in the right lobe of the liver.
December 1989
DOPPLER STUDIES OF HEPATOCELLULAR CARCINOMA
1493
Figure 3. Hepatocellular carcinoma 4.0 cm in diameter. A. A hepatocellular carcinoma in the posterior inferior subsegment of the right lobe is visualized as a hyperechoic area by US B-image (arrow]. The white bar shows the position of the lo-mm sample volume. B. Arterial Doppler signal of 1.5 kHz from within the tumor. C. Right hepatic arteriogram showing a vascular mass (arrow) in the right lobe of the liver.
1494
OHNISHI AND NOMURA
GASTROENTEROLOGY Vol. 97, No. 6
Figure 4. Hepatocellular carcinoma 7 cm in diameter. A. A hepatocellular carcinoma in the medial subsegment of the left lobe is visualized as a heterogeneous area by US Bimage. The white bar shows the position of the 2-mm sample volume. B. Arterial Doppler signal of 4.2 kHz from within the tumor. C. Proper hepatic arteriogram showing a highly vascular mass (arrow).
December 1989
DOPPLERSTUDIESOF HEPATOCELLULAR CARCINOMA 1495
Table 1. Ultrasound, Doppler, and Angiographic Characteristics of Hepatocellular Carcinomas Hepatocellular Group Mean diameter (range] (cm) B-image of hepatocellular carcinoma Hypoechoic Isoechoic with low echoic periphery Hyperechoic Heterogeneous Arterial Doppler signals obtained Within the tumor At the periphery of the tumor Within or at the periphery of the tumor or both sites Mean peak systolic frequency of arterial Doppler signal (range) (kHz1 Mean vascularity (range) Arterioportal
shunting
53 cm (n = 37) 1.9 (1.0-3.0)
carcinomas >3 cm (n = 26) 4.9 (3.2-8.7)
27 (73%) 4 (11%)
5 (19%) 0 (0%)
3 (8%) 3 (8%)
6 (23%) 15 (58%)
17(46%)
26 (70%)
20(77%) 21(81%)
28 (76%)
26 (100%)
0.7 (O-1.6)a
1.9 (0.4-4.2)a
2.7 (2-3)” n = 34 1 (3%)b
3.4 (2-4)O n = 25 8 (32%)b
Angiography was performed in 44 patients with 59 hepatocellular carcinomas. The grading of vascularity was as follows: 0 = avascular, 1 = hypovascular, 2 = isovascular, 3 = vascular, 4 = highly vascular. ’ p < 0.05. b p < 0.01.
Discussion In 1977, Wells et al. (22) reported that from among 9 patients with palpable masses in their breasts they elicited Doppler-shifted signals with a systolic modulation from 3 patients with malignancies, using an ~-MHZ continuous-wave Doppler system. They found no signals or found weak signals in the remaining 6 patients who had benign lesions. They suggested that this technique could distinguish benign breast lesions from malignant ones and that it Table 2. Ultrasound and Doppler Characteristics
might be possible to develop a new high-speed ultrasonic Doppler scanner to make breast scanning practicable. Since then breast cancer diagnosis by Doppler US has been studied by many researchers (23-27). White and Cledgett (23), Lypacewicz et al. (25), and Burns et al. (27) observed Doppler signals with systolic modulation from benign as well as malignant breast lesions. Gros et al. (24) reported that they obtained Doppler signals from every “inflammatory and evolutive lesion benign or malignant.” Regarding the spectra of the signals, in breast cancer, when lo-MHz continuous-wave Doppler units were used in particular, -90% of localized tumors exhibited abnormal properties (27). Among these were very high peak velocities, increased mean velocities, and forward and reverse flow (23-27). From these observations, it would seem that it is the quality of the Doppler signal rather than its presence or absence that differentiates malignant breast tumors from benign lesions (23,27). In 1979, Mountford and Atkinson (28) tested this technique in abdominal tumors, but found they could not obtain Doppler signals from abdominal tumors using a lo-MHz continuous-wave Doppler device. They concluded that this was because the high frequency of the ultrasound limited the penetration to 30 mm because most of the energy was attenuated in the abdominal wall. Recent technological developments in the duplex US and pulsed Doppler systems have made possible the noninvasive measurement of flow velocity and blood flow in deep abdominal and pelvic vessels (16). In 1985, Taylor and Burns (16) published the first report on obtaining abnormal signals using pulsed Doppler instruments operating at a center frequency of 3.0 MHz from several different malignant abdominal tumors, including hepatoblastoma, metastatic neuroblastoma, a mixed Mullerian cell tumor of the uterus, and a renal carcinoma. Subsequently, in 1987, Taylor et al. (17) reported the
of Isovascular,
Vascular,
and Highly Vascular
Hepatocellular
Carcinomas Hepatocellular
Group Mean diameter (range) (cm) Arterial Doppler signals obtained Within the tumor At the periphery of the tumor Within or at the periphery of tumor or both sites Mean peak systolic frequency of arterial Doppler signal (range) (kHz) Arterioportal shunting
Isovascular (n = 12)
carcinomas
Vascular (n = 35)
Highly vascular (n = 12)
2.0 (1.0-3.6)
2.7 (1.0-6.0)
5.7 (4.0-8.7)a.b
4 (33%) 5 (42%) 6 (50%)
17 (49%) 27 (77%) 32 (91%)
11(92%) 12 (100%)
0.5 (O-1.6)
1.0 (O-3.1)
2.5 (0.7-l.2)",b
0 (0%)
2 (6%)
11(92%)
7 (58%)
Angiography was performed in 44 patients with 59 hepatocellular carcinomas. The grading of vascularity was as follows: 0 = avascular, 1 = hypovascular, 2 = isovascular, 3 = vascular, 4 = highly vascular. 0 p < 0.05 vs. isovascular group. b p < 0.05 vs. vascular group.
1496
OHNISHI AND NOMURA
GASTROENTEROLOGY Vol. 97, No. 6
periphery of hepatocellular carcinomas, other hepatic focal lesions, and hepatic areas other than focal lesions? (b) Are the Doppler signals associated with hepatocellular carcinoma different from those from the other hepatic focal lesions or those from the other parts of the liver? The present study demonstrated that arterial Doppler signals were obtained either from within the body, at the periphery, or from both regions in about 75% of small hepatocellular carcinomas and in all the larger carcinomas, Arterial Doppler signals were obtained only at the periphery of some cholangiocarcinemas, some liver metastatic tumors, and some hemangiomas, whereas they were not obtained at all from regenerative nodules, pseudotumors, and liver cysts. This result seemed to be in accord with the angiographic findings of these hepatic focal lesions. Most hepatocellular carcinomas are vascular both within the body and at the periphery. Hemangiomas, cholangiocarcinomas, and metastases are more vascular at the periphery. Regenerative nodules and pseudotumors are isovascular, whereas liver cysts
Figure 5. Regenerative nodule 2.0 cm in diameter. A regenerative nodule in the medial subsegment of the left lobe is visualized as a hypoechoic area by US B-image (arrows]. The white bar shows the position of the z-mm sample volume.
results of Doppler examination in 68 consecutive patients with focal liver lesions, including 12 hepatocellular carcinomas that were all >4 cm in diameter. All 10 tumors with Doppler shifts of ~5 kHz proved to be hepatocellular carcinomas, and the correlation with angiographic findings suggested that the high-velocity Doppler signals were associated with large pressure gradients due to arterioportal shunting. These reports encouraged us to determine whether duplex Doppler US was also useful for the diagnosis of small hepatocellular carcinomas 53 cm in diameter as well as for tumors with a diameter >3 cm. Taylor et al. did not state whether Doppler signals were obtained from within the tumors or at their peripheries because most of the carcinomas they studied were diffuse or multifocal. However, we believed that it might be important for the diagnosis of small hepatocellular carcinomas to determine the precise origin of the Doppler signal. Therefore, we investigated two main areas. (a) Can a pulsed Doppler device operating at a center frequency of 3.0 MHz detect Doppler signals from within or at the
Figure 6. Pseudotumor 1.8 cm in diameter. A pseudotumor is ... . me .. liver .. ’ ’ area wifnin visualized as a h!qpoecnoic parenchyma with diffl use increased echogenicity by US Ble White image (arrow). Tl._ .._____bar shows the position of the Z-mm sample volume.
December 1989
are avascular. No Doppler signals were obtained from the hepatic parenchyma other than from the focal lesions, which differed somewhat from the findings of Taylor et al. that Doppler signals of 50.7 kHz were obtainable from the hepatic parenchyma (I 7). We found that the mean peak systolic frequency for hepatocellular carcinoma was significantly greater than for cholangiocarcinoma and hemangioma, which was in accordance with the report of Taylor et al. (17). The mean peak systolic frequency of the hepatocellular carcinomas was related to their size, vascularity, and degree of arterioportal shunting, with the peak frequency being significantly less in smaller tumors. A peak systolic frequency of 23 kHz was obtained only in hepatocellular carcinomas 24 cm in diameter with arterioportal shunting and the peak frequency was C2.6 kHz in smaller carcinomas and in other hepatic focal lesions. These findings led us to conclude that duplex Doppler US is of limited value in the diagnosis of hepatic focal lesions including small hepatocellular carcinomas. When arterial Doppler signals are obtained from within the body of a hepatic focal lesion it is probably a hepatocellular carcinoma, and when the signals are obtained at the periphery it is likely to be a tumor. However, the absence of arterial Doppler signals does not necessarily mean that the lesion is not a tumor. Measurement of the peak systolic frequency is thus of little value in the differentiation of small hepatocellular carcinomas from other lesions encountered clinically.
DOPPLER STUDIES OF HEPATOCELLULAR CARCINOMA
11.
12.
13.
14.
15. 16. 17.
18.
19.
20.
21.
22.
References 1.
Sheu JC, Sung JL, Chen DS, et al. Ultrasonography of small hepatic tumors using high resolution linear array real-time instruments. Radiology 1984;150:797-802. 2. Shinagawa T, Ohto M, Kimura K, et al. Diagnostic and clinical features of small hepatocellular carcinoma with emphasis on the utility of real-time ultrasonography: a study of 51 patients. Gastroenterology 1984;86:496-502. 3. Okuda K, Ohtsuki T, Obata H, et al. Natural history of hepatocellular carcinoma and prognosis in relation to treatment. Study in 850 patients. Cancer 1985;56:918-28. 4. Ohnishi K, Tanabe Y, Ryu M, et al. Prognosis of hepatocellular carcinoma smaller than 5 cm in relation to treatment: study of 100 patients. Hepatology 1987;7:1285-90. 5. Yates CK, Streight R. Focal fatty infiltration of the liver simulating metastatic diseases. Radiology 1986;159:83-4. 6. Kawashima A, Suehiro S, Murayama S, Russell WJ. Focal fatty infiltration of the liver mimicking a tumor: sonographic and CT features. J Comput Assisted Tomogr 1986;10:329-31. 7. Berland LL. Focal areas of decreased echogenicity in the liver at the porta hepatis. J Ultrasound Med 1986:5:157-g. 8. White EM, Simeone JF, Mueller PR, Grant EG, Choyke PL, Zeman RK. Focal periportal sparing in hepatic fatty infiltration: a cause of hepatic pseudomass on US. Radiology 1987;162:57-9. 9. Bree RL, Schwab RE, Neiman HL. Solitary echogenic spot in the liver: is it diagnostic of a hemangioma? Am J Roentgen01 1983;140:41-5. HA. Benign epithelial tumors and tumorlike 10. Edmondson
23. 24.
1497
lesions of the liver. In: Okuda K, Peters RL, eds. Hepatocellular carcinoma. New York: Wiley, 1976:309-30. Sasaki K. Adenomatous hyperplasia in liver cirrhosis: an approach from a microangiographical point of view. Gann Monogr Cancer Res 1980;25:127-40. Makuuchi M, Hasegawa H, Yamazaki S. Development of liver surgery by operative sonography. Semin Gastroenterol Surg 1983;13:151-73 (in Japanese). Akimoto S, Saito A, Watayoh T, Isobe Y, Takasaki T. Diagnostic capability and limitation of ultrasonography in small hepatocellular carcinoma. Diagn Imaging Abdomen 1984; 4:209-16 (in Japanese]. Arakawa M, Kage M, Sugihara S, Nakashima T, Suenaga M, Okuda K. Emergence of malignant lesions within an adenomatous hyperplastic nodule in a cirrhotic liver. Observations in five cases. Gastroenterology 1986;91:198-208. Itai Y, Ohnishi S, Ohtomo K, et al. Regenerative nodules of liver cirrhosis: MR imaging. Radiology 1987;165:419-23. Taylor KJW, Burns PN. Duplex Doppler scanning in the pelvis and abdomen. Ultrasound Med Biol 1985;11:643-58. Taylor KJW, Ramos I, Morse SS, Fortune KL, Hammers L, Taylor CR. Focal liver masses: differential diagnosis with pulsed Doppler US. Radiology 1987;164:643-7. Yamasaki S, Hasegawa H, Makuuchi M. Clinicopathological observation of minute liver cancer and a new method of hepatectomy: analysis of 27 resected cases. Acta Hepatol Jpn 1981;22:1714-24 (in Japanese). Takayasu K, Shima Y, Muramatsu Y, et al. Angiography of small hepatocellular carcinoma: analysis of 105 resected tumors. Am J Roentgen01 1986;147:525-9. Nakayama T, Hiyama Y, Ohnishi K, et al. Arterioportal shunts on dynamic computed tomography. Am J Roentgen01 1983; 140:953-7. Ohnishi K, Okuda K, Ohtsuki T, et al. Formation of hilar collaterals or cavernous transformation after portal vein obstruction by hepato-cellular carcinoma. Observations in ten patients. Gastroenterology 1984;87:1156-3. Wells PNT, Halliwell M, Skidmore R, Webb AJ, Woodcock JP. Tumor detection by ultrasonic Doppler blood-flow signals. Ultrasonics 1977;15:231-2. White DN, Cledgett PR. Breast carcinoma detection by ultrasonic Doppler signals. Ultrasound Med Biol 1978;4:329-35. Gros Ch M, Dale G, Gairand B. Breast echography: criteria of malignancy and results. In: Kurjak A, ed. Recent advances in ultrasound diagnosis. Amsterdam: Excerpta Medica, 1978:
292-8. 25. Lypacewicz
G, Powalowski T, Lukawska K. Ultrasonic examination of breast tumors with Doppler method. In: Fillipczynski L, Zieniuk JK, eds. Proceedings of the Second Congress of the Federation of Acoustical Societies of Europe. Volume II. Polish Academy of Sciences, 1978:153-6. assessment of an 26. Minasian H, Bamber JC. A preliminary ultrasonic Doppler method for the study of blood flow in human breast cancer. Ultrasound Med Biol 1982;8:357-64. 27. Burns PN, Halliwell M, Wells PNT. Ultrasonic Doppler studies of the breast. Ultrasound Med Biol 1982;8:127-43. 28. Mountford RA, Atkinson P. Doppler ultrasound examination of pathologically enlarged lymph nodes. Br J Radio1 1979; 52:464-7.
Received February 17, 1988. Accepted June 2, 1989. Address requests for reprints to: Kunihiko Ohnishi, M.D., First Department of Medicine, Chiba University School of Medicine, 1-8-1 Inohana-cho, Chiba City, Chiba 280, Japan. This study was supported in part by a research grant from the Intractable Diseases Division of the Public Health Bureau, Ministry of Health and Welfare, Japan.