ª Springer Science+Business Media New York 2015

Abdominal Imaging

Abdom Imaging (2015) DOI: 10.1007/s00261-015-0406-z

Comparison of a pocket-size ultrasound device with a premium ultrasound machine: diagnostic value and time required in bedside ultrasound examination Konrad Friedrich Stock,1 Bettina Klein,1 Dominik Steubl,1 Christian Lersch,2 Uwe Heemann,1 Stefan Wagenpfeil,3 Florian Eyer,4 Dir-Andre Clevert5 1

Abteilung fu¨r Nephrologie, Technische Universita¨t Mu¨nchen, Munich, Bavaria, Germany Medizinische Klinik und Poliklinik, Technische Universita¨t Mu¨nchen, Munich, Germany 3 Institut fu¨r medizinische Statistik und Epidemiologie, Technische Universita¨t Mu¨nchen, Munich, Germany 4 Abteilung fu¨r Toxikologie, Technische Universita¨t Mu¨nchen, Munich, Germany 5 Institut fu¨r Radiologie, Klinikum der Universita¨t Mu¨nchen, Munich, Germany 2

Abstract Purpose: Time savings and clinical accuracy of a new miniature ultrasound device was investigated utilizing comparison with conventional high-end ultrasound instruments. Our objective was to determine appropriate usage and limitations of this diagnostic tool in internal medicine. Methods: We investigated 28 patients from the internalmedicine department. Patients were examined with the Acuson P10 portable device and a Sonoline Antares instrument in a cross-over design. All investigations were carried out at the bedside; the results were entered on a standardized report form. The time for the ultrasound examination (transfer time, setting up and disassembly, switching on and off, and complete investigation time) was recorded separately. Results: Mean time for overall examination per patient with the portable ultrasound device was shorter (25.0 ± 4.5 min) than with the high-end machine (29.4 ± 4.4 min; p < 0.001). When measuring the size of liver, spleen, and kidneys, the values obtained differed significantly between portable device and the high-end instrument. In our study, we identified 113 pathological ultrasound findings with the high-end ultrasound machine, while 82 pathological findings (73%) were concordantly detected with the portable ultrasound device. The main diagnostic strengths of the portable device

Correspondence to: Dominik Steubl; email: [email protected]

were in the detection of ascites (sensitivity 80%), diagnosis of fatty liver, and identification of severe parenchymal liver damage. Conclusions: The clinical utility of portable ultrasound machines is limited. There will be clinical roles for distinct clinical questions such as detection of ascites or pleural effusion when used by experienced examiners. However, sensitivity in detecting multiple pathologies is not comparable to high-end ultrasound machines. Key words: Pocket-size ultrasound—Abdominal ultrasound—Emergency ultrasound—Examination time—Handheld ultrasound

Portable ultrasound devices have been commercially available for several years now, and their value for clinical use has already been the subject of several studies. Among the roles investigated are time saving [1] and the assessment of specific clinical questions [2, 3]. Other studies have used pocket-size ultrasound devices to supplement computed tomography [4] or have compared the clinical value of portable ultrasound devices with that of high-end instruments [5–7]. Several authors have evaluated the role of portable ultrasound devices in intensive medicine [8, 9] and in pre-hospital emergency medicine [10–12]. Scientific evaluation of portable devices is not yet complete, as technical development of ultrasound devices proceeds apace. Presently, miniaturized ultrasound de-

K. F. Stock et al.: Comparison of a pocket-size ultrasound device with a premium ultrasound machine

Fig. 2. Intrahepatic biliary duct. Enlargement diagnosed with the Acuson P10.

Fig. 1.

Acuson P10 miniature ultrasound device.

vices reduced to pocket size are available. The vision of an ‘‘ultrasound stethoscope,’’ as anticipated by the authors of the pilot studies on portable ultrasound devices [13], now seems the reality. Our index machine for comparison fits into a lab coat pocket. Its use is limited to the battery, which needs to be recharged after a few hours. While this may open up new avenues for the use for portable instruments as an extension to traditional physical medical examinations (Fig. 1), conversely, increasing miniaturization may compromise image quality subsequently reducing diagnostic accuracy of perception and interpretation if inexperience and a lack of commitment to quality prevail; such that this may not only increase clinical examination time, but also reduce its appropriate utility. Any demonstrable time saving, achieved by the use of portable devices (shorter boot time, quicker transfer, and easier positioning at the patient’s bedside) [1], has to be counter-balanced by the awareness of limitations of use of the miniature device in both a- and post priori situations. More detailed enunciation, by obtaining empirical data to consider rationally appropriate applications of the device in internal medicine given these parameters, is desirable and is the aim of this study.

Materials and methods We prospectively investigated 28 randomly invited inpatients in the departments of nephrology, gastroenterology, and toxicology over a period of 2 weeks. This prospective pilot study was approved by the ethics committee of the Technical University of Munich and

conducted in accordance with its guidelines. Informed consent was obtained from all patients. The age range of the patients was 24–94 years (mean 68 years); 12 were male and 16 female. The mean BMI of the patients was 25 (standard deviation ± 4). Examinations were carried out according to a standardized protocol by two internal-medicine specialists experienced in ultrasonography (more than 10 years, both board-certified instructors of the National ultrasound society). The patients were first examined with the ‘‘Acuson P10’’ pocket ultrasound device (=PUD, 2– 4 MHz, phased array transducer, B mode; Figs. 1, 2). A second bedside examination of the patients with a highend instrument ‘‘Sonoline Antares’’ (=HEI, 2–6 MHz, curved array; B mode only; one physician performing the examination, the second closely observing the examination on the monitor). The examiners had experience in the use of both machines (training period of one month on the PUD before initiation of the study). The results were discussed and agreed conclusion was obtained after the end of the exam (‘‘gold standard’’ of the exam). Examinations with PUD were randomly performed by one of the examiners at random; the subsequent examination of the same patient using the HEI was then performed by the other examiner with the initial ‘‘PUD’’ examiner monitoring and documenting the examination using HEI. This methodology prevented the examiner from being previously primed with the patient’s pathologies and sonographic windows. Each of the examiner performed 14 PUD and 14 HEI examinations. We did not use computed tomography or magnetic resonance imaging as the gold standard comparison technique since the aim of the study was to compare ultrasound devices and not necessarily detect all pathologies. Both examinations were carried out on the ward, at the bedside. The gastroenterology, nephrology, and toxicology wards can be at a considerable distance away

K. F. Stock et al.: Comparison of a pocket-size ultrasound device with a premium ultrasound machine

Table 1. Pathological findings with the Acuson P10 and the high-end Sonoline Antares instrument Pathological findings Total Liver, total Focal liver lesions Severe parenchymal damage Fatty liver Hepatic venous congestion Gallbladder and biliary system Gallstones Gallbladder-wall thickening Hydrocholecystis Gallbladder polyp Intrahepatic cholestasis Stents in DHC Dilated DHC Aerobilia Kidney, total Cystic compressive processes Hydronephrosis Pancreatic disease Aortic sclerosis Pleural effusion Ascites Additional findings Splenic infarction Gastroparesis

n

Diagnoses that agree n

P10 false-neg. n

P10 sensitivity (%)

P10 specificity (%)

113 38 11 5 20 2 22 7 2 2 1 3 2 4 1 15 14 1 2 8 21 5 2 1 1

82 32 5 5 20 2 13 4 0 1 1 3 2 2 0 12 11 1 1 5 13 4 2 1 1

31 6 6 0 0 0 9 3 2 1 0 0 0 2 1 3 3 0 1 3 8 1 0 0 0

73 84 45

99 100 100

59 57

100 100

80 79

100 100

63 62 80

100 89 100

from the ultrasound department. For each examination, the findings were documented by an assistant on a standardized report form (22 subitems; ultrasound regions: abdomen, pleura). With the aid of this report form, parametric variables (size of liver, kidneys, spleen, abdominal aorta, vena cava, and common bile duct diameter) were determined. In addition, patients were screened for certain diseases, as well as for any additional findings (Table 1). The ultrasound images were stored digitally during the examination on both instruments. The time required for the total examination process was likewise measured, in each case, by the assistant, using a stopwatch. The times required to fetch the instrument and take it away, to set it up and disassemble it, for booting, and for the performance of the clinical Bmode examination were determined. We also ascertained the times needed to measure the size of organs with both instruments. For our study, the high-end instrument was started from standby mode; however, we also, separately, measured the times for starting the ultrasound instrument from ‘‘off’’ which, during night hours, is the most likely routine scenario.

Statistics The quantitative comparisons for the analysis of times and measured data were made with the t test for two independent samples. All the p values are two sided and based on a significance level of 0.05. The evaluation was carried out with SPSS version 15.0, data management with Excel 2011.

Fig. 3. Comparison of examination times using Acuson, P10 and Acuson Antares.

Results The time required for the total individual examination process with the portable ultrasound device was 25.0 (±4.5) min, which was shorter than with the high-end machine (29.4 ± 4.4 min, p < 0.001) when the measurement was made without using the standby function but started from complete shut-down situation (Fig. 3). Apart from the shorter transfer time, the decisive factor here was the shorter set-up and disassembly and boot time with the PUD (mean time required is 6.6 min with the PUD as opposed to 16.5 min with the HEI). In contrast, the mean time required for the examination process when the HEI was used on standby was 24.5

K. F. Stock et al.: Comparison of a pocket-size ultrasound device with a premium ultrasound machine

Fig. 4.

Comparison of hands-on times.

(±4.4) min, making the mobile high-end instrument somewhat faster even than the portable system. With regard to the hands-on time (time needed for the interactive imaging component examination only), the use of the PUD was associated with a longer time than with the HEI (18.1 vs. 13.2 min, Fig. 4). The mean time required for the measurement of organ sizes was 4.5 min with the portable system as opposed to 2.0 min with the high-end instrument (these values were determined in 21 patients). The imaging time without measurement of organ sizes is likewise greater with PUD (13.6 min) than with HEI (11.2 min). Organ size measurements differed; the values determined for the size of the liver were significantly smaller with the PUD (14.9 ± 1.9 cm) than with the HEI (16.8 ± 2.9 cm, p < 0.001), the difference being 1.9 cm. The same applied to the size of the spleen (10.8 ± 1.9 cm with PUD as opposed to 11.2 ± 2.1 cm with the HEI, p = 0.008), on average measured as 0.4 cm smaller with the PUD. Accordingly, the PUD correctly identified only 8 (44%) of the 18 cases of hepatomegaly and only 6 (75%) of the 8 cases of splenomegaly that had been seen with the high-end instrument. There were no false-positive findings. Renal size values obtained with the PUD (9.8 ± 2.2 cm) likewise differed substantially from those measured with the HEI (10.4 ± 1.5 cm, p < 0.001). Kidney size with the pocket ultrasound device was thus measured on average 0.6 cm smaller. The PUD correctly detected 4 out of the 4 cases of atrophic kidneys. The PUD wrongly identified ten kidneys as atrophic, however, which represents a false-positive rate of 20%. For aorta, there were statistically but not clinically relevant differences between the lumen diameters measured with the PUD and those measured with the HEI (1.6 ± 0.3 vs. 1.7 ± 0.3 cm, p = 0.03). A similar picture was observed with the vena cava (1.7 ± 0.4 cm with the PUD vs. 1.9 ± 0.5 cm with the HEI, p = 0.01). There were neither statistically nor clinically relevant CBD diameter measurement differences between the PUD and the HEI (4.7 ± 1.4 vs. 4.8 ± 1.4 mm; p = 0.8). A total of 113 pathological findings in 28 patients were identified with the high-end instrument; 82 (73%) of

these findings were also correctly seen with the portable device (Table 1). The portable device detected 32 (84%) of the 38 pathological liver findings, but only 5 (45%) of the 11 hepatic compressive processes. All cases of severe liver parenchymal damage due to liver cirrhosis (5 cases), fatty liver (20 cases), and hepatic venous congestion (2 cases) were detected with the pocket ultrasound device. Of the 22 gallbladder and biliary-system pathological morphologies identified by the high-end instrument, 13 (59%) were also detected with the portable device. 4 (57%) of the 7 cases of gallstones were detected with the PUD, but neither of the 2 cases of gallbladder-wall thickening. A case of gallbladder polyp was correctly identified by the PUD, as was 1 of the 2 cases of hydrocholecystis. Pronounced intrahepatic biliary duct dilatation was present in 3 patients in the group and was picked up by the PUD in all cases. Stents in the common biliary duct (2 cases) were likewise correctly identified with the PUD. The PUD detected 80% (12) of the 15 pathological morphologies in the kidneys: 11 (79%) of the 14 renal compressive processes were picked up by the PUD, as was one case of hydronephrosis. The PUD picked up 1 of the 2 cases of pathological morphologies in the pancreas detected by the high-end instrument. Sclerosis of the abdominal aorta was identified by the PUD in 63% of cases. At 62%, its sensitivity for the detection of pleural effusion was likewise low. In contrast, the PUD detected 80% (4 out of 5) of the cases of ascites.

Discussion It was possible to achieve shorter examination times with the pocket ultrasound device than with the conventional high-end instrument due to the short boot time and rapid set-up and disassembly, and also the time saved on the transfer of the portable device. However, the actual scan time (B-mode images and measurement times) was significantly shorter with the high-end ultrasound machine. The difference in the time required for transfer and set-up (set-up, disassembly, and boot time), which on average takes 6.6 min for the PUD and 16.5 min for the HEI, is obvious. Less pronounced, though still significant, was the difference in the times required for the total examination process (mean time of 25 min with the portable device as opposed to 29.4 min with the high-end instrument) for a standardized abdominal exam (following our protocol). These results are comparable to the conclusions of other studies on the time saving with portable ultrasound devices [1]. Through comparing a standardized examination protocol and procedure for every patient, it was possible for the first time to make a comparison between the actual hands-on times required in each case to obtain the B-mode image. A clear difference was found: 18.1 min with the PUD as opposed to 13.2 min with the HEI. We initially assumed that the greater time

K. F. Stock et al.: Comparison of a pocket-size ultrasound device with a premium ultrasound machine

required with the PUD might be due solely to longer times taken for the subjectively somewhat more laborious process of measuring organ sizes. This was not confirmed however. With the PUD, measuring organ sizes took, on average, 2.5 min longer—and B-mode imaging without measurement of organ sizes 2.4 min longer—than with the HEI. We attribute this greater time requirement to the smaller size and poorer resolution of the monitor, which places greater demands on the investigator’s concentration and eyesight and lengthens the time to definitive diagnosis. However, the shortness of the period of experience of operation of the mini device could also lead to delays which might be reduced with repetitive, routine use of the PUD. We also acknowledge that usually the main clinical use of the PUD is not to perform a complete abdominal ultrasound examination but focus on special clinical ‘‘triage microquestions,’’ e.g., is ascites/pleural effusion/enlargement of the intra- and extra-hepatic biliary duct system present? In the determination of organ sizes (liver, spleen, and kidneys), the measurements obtained with the PUD differed systematically from those with the high-end instrument. These common cause differences were most pronounced in the determination of the size of the liver, which on average was 1.9 cm smaller with the PUD than with the HEI. This also explains why hepatomegaly was correctly identified in only 44% of the cases by the portable device. The spleen measurements were likewise smaller with the PUD. On average, these were only 0.4 cm smaller than with the high-end instrument, however. As a result, the PUD was less sensitive in the identification of splenomegaly. This condition was correctly detected in 75% of the cases. The kidney measurements were on average 0.6 cm smaller with the PUD; 100% of the cases of atrophic kidneys were correctly identified. There were also false-positive results, to a rate of 20%. This type-1 error can be readily explained by the smaller measurements obtained with the PUD. To sum up, organ size measurements are substantially smaller than with the portable instrument, which can make diagnosis (particularly diagnosis of borderline hepatomegaly and splenomegaly) less reliable leading to kidneys being wrongly assessed as atrophic. With the high-end instrument, hepatomegaly is considered to be present if the liver length measured is 16 cm or over, splenomegaly is considered to be present if the spleen length measured is 12 cm or over, and atrophic kidney is considered to be present if the kidney length is smaller than 9 cm, though these normal values are debatable varying from and within different centers. The values obtained for liver, spleen, and kidney size measurements should be reconsidered in regards to the use of the PUD, and these differences need to be taken into account in the assessment of organ sizes. This variation of measure-

ments might be explained by the different transducers used and their underlying physical properties. The deterioration in lateral resolution in the B-mode picture of the phased array transducer (PUD) with increasing depth due to reduced ultrasound line density might well explain this finding. Although there was in some instances a statistically relevant difference between the values measured with the PUD and the HEI for the aorta and vena cava and for the determination of CBD diameter, the difference was not clinically relevant. In the course of ultrasonographic examination of the 28 patients from our internal-medicine department, we detected a total of 113 pathological findings. The PUD detected 82 (73%) of the pathological findings detected with the high-end instrument. The strengths of the portable device were particularly in the detection of ascites (sensitivity 80%) and of liver parenchyma damage from liver cirrhosis or fatty liver. As expected, the diagnostic evaluation of focal liver lesions was a weak point for the PUD. Definitively, this should not be undertaken with a miniature device. Exclusion of gallstones was also too unreliable with the miniaturized device. Surprisingly, its sensitivity for the diagnosis of pleural effusion was only 62%, here also being false-positive findings. This is attributed to the occurrence of artifacts and their incorrect assessment with the miniaturized device as well as due to the limited ultrasound field of the PUD. In the kidney, cystic compressive processes were readily detected, and a single case of hydronephrosis was also identified with PUD. The PUD is probably particularly suitable for the detection of hydronephrosis; no statistically relevant conclusions in this regard can be drawn from our group of patients and a larger study would be required.

Conclusion Although the acquisition costs for the PUD are moderate, the results in this study suggest that it is of limited clinical utility: you get what you pay for. The slight advantage of PUD examination time does not compensate its underperformance in diagnostic accuracy. Due to the limited quality of the B-mode picture, resulting from the phased array transducer, PUD should only be used by very experienced examiners. Even then the sensitivity for various pathologies, especially those requiring a high-resolution B-mode picture such as focal lesions in solid organs, is prohibitively limited. Although we did not document a high number of false-positive findings, we recommend all pathologies documented with PUD to be re-evaluated with departmental examination—thus doubling the effort which would have to be counter-balanced by bed usage cost savings. Clinical use of PUD should be limited to

K. F. Stock et al.: Comparison of a pocket-size ultrasound device with a premium ultrasound machine

answering specific clinical questions, e.g., arising in the emergency room such as is there ascites or enlargement of the biliary duct system present? The use of a PUD to answer such sporadic individual clinical questions is not unrestricted due to the limited battery life, yet still allowing sufficiency in numbers of patients who may be appropriately examined on ward (take-) rounds. Nevertheless, the considerable limitations in the scope and numbers of cases appropriate to responsible use of the miniature device must always be foremost in mind. Before employing PUD in clinical practice, it will be important to robustly test its accuracy, utility, and cost effectiveness in future studies of limited point-of-care questions. Acknowledgments. The authors thank Dr. Tom Fitzgerald, The Royal Infirmary of Edinburgh, for the critical review and proof-reading of the manuscript.

References 1. Fischer T, Filimonow S, Petersein J, et al. (2002) Ultrasound at the bedside: does a portable ultrasound device save time? Ultraschall Med 23(5):311–314 2. Vourvouri EC, Poldermans D, Schinkel AFL, et al. (2001) Abdominal aortic aneurysm screening using a hand-held ultrasound device. ‘‘A pilot study’’. Eur J Vasc Endovasc Surg 22(4):352–354

3. Ryan SM, Smith E, Sidhu PS (2002) Comparison of the SonoSite and Acuson 128/XP10 ultrasound machines in the ‘bed-side’ assessment of the post liver transplant patient. Eur J Ultrasound 15(1–2):37–43 4. Oschatz E, Prosch H, Schober E, Mostbeck G (2004) Evaluation of a portable ultrasound device immediately after spiral computed tomography. Ultraschall Med 25(6):433–437 5. Seitz K, Vasilakis D, Ziegler M (2003) Efficiency of a portable Bscan ultrasound device in comparison to a high-end machine in abdominal ultrasound. Results of a pilot study. Ultraschall Med 24(2):96–100 6. Ziegler CM, Seitz K, Leicht-Biener U, Mauch M (2004) Detection of therapeutically relevant diagnoses made by sonography of the upper abdomen: portable versus high-end sonographic units—a prospective study. Ultraschall Med 25(6):428–432 7. Pieckenpack A, Klebl F, Dorenbeck U, et al. (2002) Evaluation of a new, portable ultrasound system in routine clinical use. Rofo 174(3):349–352 8. Beaulieu Y (2007) Bedside echocardiography in the assessment of the critically ill. Crit Care Med 35(5 Suppl):S235–S249 9. Beaulieu Y, Marik PE (2005) Bedside ultrasonography in the ICU: part 2. Chest 128(3):1766–1781 10. Price DD, Wilson SR, Murphy TG (2000) Trauma ultrasound feasibility during helicopter transport. Air Med J 19(4):144–146 11. Busch M (2006) Portable ultrasound in pre-hospital emergencies: a feasibility study. Acta Anaesthesiol Scand 50(6):754–758 12. Lapostolle F, Petrovic T, Lenoir G, et al. (2006) Usefulness of hand-held ultrasound devices in out-of-hospital diagnosis performed by emergency physicians. Am J Emerg Med 24(2):237–242 13. Judmaier G, Seitz K (2004) How reliable is sonography of the upper abdomen with portable sonographic units? What does the future hold? Ultraschall Med 25(6):408–410