Sonographic Findings in Cases ofMissed Gallstones

Kedar N. Chintapalli, MD, Abraham A. Ghiatas, MD,* Shailendra Chopra, MD, MRCP, FRCR,Beatrice Escobar, MD, Christine C. Esola, MD, Gerald D. Dodd III, MD

Department of Radiology, The University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Drive,San Antonio, Texas 78284-7800

Received 20 April 1998; accepted 29 September 1998

ABSTRACT: Purpose. We retrospectively evaluatedsonographic findings in 946 cases of gallstones to de-termine whether the false-negative rate for gallstonedetection by sonography has decreased as a result oftechnologic advances over the past 15 years.

Methods. We reviewed preoperative sonographicreports, operative notes, and pathologic reports for614 women and 332 men (ages 22–78 years) seen overa 2.5-year period and compared sonographic findingswith surgical pathologic findings after cholecystecto-my. Sonograms for patients whose gallstones weremissed on sonography were reviewed by 3 board-certified radiologists.

Results. Preoperative sonography of the gallblad-der accurately predicted the presence of gallstones in934 cases (98.7%). Gallstones were not identified bysonography in the remaining 12 cases. In those cases,sonography revealed polyps in 5, sludge in 5, sludgeplus a polyp in 1, and neither stones nor polyps in 1.Thus, the false-negative rate was 1.3%.

Conclusions. Despite improvements in sono-graphic technology, detection of small gallstones re-mains difficult in some cases. Adherent gallstones canmimic gallbladder polyps. Our false-negative rate fordetection of gallstones was no different from that inearlier studies. © 1999 John Wiley & Sons, Inc. J ClinUltrasound 27:117–121, 1999.

Keywords: gallstones; ultrasonography; false-nega-tive rate

Soon after its introduction, sonography of thegallbladder became the tool of choice for diag-

nosing cholelithiasis.1–3 Nevertheless, sonogra-

phy has a low but definite false-negative rate inthe detection of gallstones. Recent advances insonographic technology have improved spatialresolution. We undertook this retrospective studyto evaluate the spectrum of sonographic findingsin cases in which gallstones were missed by so-nography and to compare the false-negative ratewith that of earlier studies.

PATIENTS AND METHODS

For this retrospective review, we searched the ra-diologic and surgical databases from the 2 hospi-tals affiliated with our institution for patientswho had undergone (1) gallbladder sonographyfor either right upper-quadrant pain or abnormalliver function tests and (2) cholecystectomywithin the ensuing 2 weeks over a 2.5-year period(March 1994 to August 1996). Cases referred forsonographic evaluation for other indications, suchas abdominal pain, hepatomegaly, or hydrone-phrosis, were not included in the study. Subjectswho had eaten soon before undergoing sonogra-phy were reexamined after an overnight fast. Allcases included in the study thus had technicallyadequate sonographic examinations. The studyincluded 946 subjects, 614 women and 332 men,whose ages ranged from 22 to 78 years. The sub-jects’ medical records, including sonography re-ports, operative notes, and pathology reports,were reviewed for this study.

All preoperative sonographic examinationswere performed with either a Spectra (Diasonics,Milpitas, CA) or an XP10 (Acuson, MountainView, CA) ultrasound scanner using 2.5–5-MHzcurved-array or sector transducers. When the

Correspondence to: K. N. Chintapalli*Present address: Voutsina 1a, Ekali, Athens 14565, GreecePresented at the 26th Annual Meeting of the Society of Gas-trointestinal Radiology, Cancun, Mexico, March 9–14, 1997

© 1999 John Wiley & Sons, Inc. CCC 0091-2751/99/030117-05

VOL. 27, NO. 3, MARCH/APRIL 1999 117

gallbladder could not be optimally visualized at 3or 5 MHz, a 2.5-MHz frequency was used. All gall-bladders were imaged via a right anterior or lat-eral intercostal or subcostal approach in both thetransverse and longitudinal planes with the pa-tients in the supine and left lateral decubitus po-sitions. All sonographers had at least 5 years ofexperience. The technical adequacy of the exami-nation (ie, gain and focal zone) was evaluated atthe time of scanning by a radiology resident, anabdominal imaging fellow, or a faculty radiolo-gist, and additional images were obtained whenthe study was believed to be incomplete.

Intraluminal foci that moved with a change in

the patient’s position with or without shadowingand shadowing hyperechoic foci with layeringwere interpreted as gallstones. Nonshadowinguniform homogeneous debris that moved slowlyin the lumen was interpreted as sludge, and non-mobile, nonshadowing intraluminal foci were in-terpreted as polyps. When the contents of thegallbladder were echogenic, mobile foci in thegallbladder that were larger than the sludge par-ticles were considered to represent nonshadowinggallstones.

Preoperative sonographic findings were com-pared with operative and pathologic findings. Ifthe sonographic findings did not correlate with

FIGURE 1. Longitudinal sonograms of the gallbladder in a 23-year-old woman with abdominal pain. (A) Imageobtained with a 2.5-MHz transducer reveals an echogenic, nonshadowing focus believed to be a sludge ballor a polyp in the fundus (arrow). (B) Repeat scan with a 3.5-MHz transducer showed the same abnormality(arrow). Multiple small cholesterol stones, but no polyps, were found at surgery.

CHINTAPALLI ET AL

118 JOURNAL OF CLINICAL ULTRASOUND

the surgical findings, then the sonograms werereviewed by 3 board-certified radiologists for thepresence of stones, polyps, or masses.

Sonographic findings were considered false-negative if the retrospective review of the imagesdid not show gallstones that were found at sur-gery.

We used the chi-squared and 2-tailed Fisher’sexact test to compare our false-negative rate withrates reported by others.5–10

RESULTS

All 946 patients had gallstones in the gallbladderat surgery. In 934 cases (98.7%), preoperative so-nography showed single or multiple mobile echo-genic foci with or without acoustic shadowingwithin the gallbladder; in 11 cases, sonographicfindings indicated polyps (5 cases), sludge (5), orboth (1) (Figures 1–4); and in 1 case, sonographyshowed no abnormalities (Table 1). In all of the 12false-negative cases, the sonographic findings re-ported at the time of the sonographic examinationand the findings from the retrospective review by3 radiologists were in agreement. The size of thepolyps seen on sonograms ranged from 0.2 to 1.0cm. The amount and appearance of the sludgevaried. Gallstones were 5 mm or smaller in 10patients and less than 1.0 cm in all 12 patients.

The false-negative rate for the sonographic di-agnosis of gallstones was 1.3%. This was not sig-nificantly different from the false-negative ratespublished by others5,6,8–10 (p 4 0.4, chi-squaredtest; p 4 0.3, Fisher’s exact test) (Table 2), exceptfor 1 study done with B-mode sonographic equip-ment.7

DISCUSSION

Sonography of the gallbladder has largely re-placed oral cholecystography in the evaluation ofgallbladder disease. Several studies published inthe past 15 years have reported high sensitivityand specificity for sonography in this setting.1–10

False-negative sonographic findings can be due togallstones’ being hidden in the cystic duct or be-hind a fold, patient obesity, recent surgery, or op-erator inexperience. The spectrum of findings insuch cases includes nonshadowing foci in the gall-bladder or echogenic foci with shadowing in theregion of the gallbladder.

Crade et al11 reported the accuracy of 3 sono-graphic patterns in predicting gallstones. An in-traluminal echogenic focus with shadowing(Crade type 1) led to the correct diagnosis of cho-lelithiasis in all cases (100% accuracy). Whenthere was shadowing in the region of the gallblad-der or the gallbladder could not be seen in pa-tients who had no surgical history (Crade type 2),diagnostic accuracy was 96%. When nonshadow-ing focal opacities were present within the gall-bladder lumen (Crade type 3), the accuracy was61%. Eleven of the 12 false-negative cases in ourseries were Crade type 3. Simeone et al12 subdi-vided the nonshadowing opacities category into 3subtypes. When 1 or more nonshadowing echo-genic foci were seen (type 3A), gallstones werepresent in 82% of cases (17 of 21). In contrast,patients with sludge and a fluid–fluid level (type3B) had gallstones in only 8% of cases (1 of 12).The third pattern, ill-defined echoes in the depen-dent portion of the gallbladder (type 3C), correctlypredicted small gallstones in 40% of cases (2 of 5).

FIGURE 2. Longitudinal sonogram of the gallbladder in a 30-year-oldwoman with abnormal liver function tests. Image obtained with a3.5-MHz transducer shows a nonmobile echogenic focus believed tobe a polyp (arrow). This was found at surgery to be a stone.

FIGURE 3. Longitudinal sonogram of the gallbladder in a 73-year-oldman with right upper-quadrant pain. Image obtained with a 3.5-MHztransducer shows only a small polyp in the anterior wall (arrow). Apolyp and a 0.8-cm stone were found at surgery.

MISSED GALLSTONES

VOL. 27, NO. 3, MARCH/APRIL 1999 119

Most of the stones in that series were smallerthan 5 mm. In our series, 6 of the 12 false-nega-tive cases were type 3A, 5 were type 3B, and theremaining 1 appeared normal. At surgery, 5 of the6 type 3A cases (interpreted as polyps) turned outto be adherent stones, and only 1 was a polyp.Stones were found in all 5 type 3B cases (inter-preted as sludge) at surgery.

Published false-negative rates 5–10 have rangedfrom 1.3% to 19.4%; this variation is due in part tothe small numbers of subjects with negative sono-grams who underwent surgery. The study withthe highest rate was an older study that involvedB-mode sonography.7 When that study is ex-cluded, the false-negative rates are between 1%and 4%.

We examined our data to determine whether alower-frequency transducer was used in cases inwhich gallstones were missed. A 3.5-MHz fre-quency was used for 11 of the 12 false-negatives,including the case with no sonographic abnor-malities, and a 4-MHz frequency was used for the12th. When a low-frequency transducer (3.5 MHz

or 2.25 MHz) must be used to visualize the gall-bladder in a large patient, electronic beam focus-ing may allow optimization of the focal zone, butthe resolution may still be limited owing to thefrequency of the transducer.

Acoustic shadowing from intraluminal foci is acharacteristic of gallstones. One of the factorsthat influence shadowing is the transducer fre-quency. Ex vivo studies have demonstrated thatcalculi larger than 0.3 cm produce acoustic shad-owing regardless of their composition.13,14 In astudy by Grossman,15 use of a 5-MHz transducerrevealed 0.1 cm stones to be echogenic withoutshadowing and 0.2–0.3-cm stones to have shad-owing; use of a 2.25-MHz transducer revealed 0.4-cm stones to have shadowing. However, in clinicalpractice, other factors determine the selection ofthe transducer, such as location of the gallblad-der, the size of the patient, and the presence offatty infiltration of the liver. The beam width alsoaffects shadowing from a stone. Filly et al,14 usinga tissue phantom, found that shadowing waspresent when the stone was at or near the centerof the beam but not if it was at the periphery.Conceivably, use of a 5-MHz transducer may im-prove detection of stones, but this may not bepractical with large patients. Simeone et al12,16

showed that individuals with sludge on sono-grams had a 1–5% chance of having small gall-stones. The practical question that remains is theclinical significance of these stones, ie, whethertheir detection would alter the management ofthe case.

FIGURE 4. Longitudinal sonogram of the gallbladder in a 40-year-old woman with right-side pain. Imageobtained with a 3.5-MHz transducer shows sludge along the posterior wall. A 0.6-cm stone and sludge werefound at surgery.

TABLE 1

Findings in 12 Sonographically False-Negative Cases

of Gallstones

Finding

No. Cases

Sonography Surgery

Normal 1 0Sludge 6 6Polyp 6 1Stones 0 12

CHINTAPALLI ET AL

120 JOURNAL OF CLINICAL ULTRASOUND

The major limitation of our study is its retro-spective nature. Only patients who underwentcholecystectomy were included; thus, some gall-stones probably went undetected in patients withnormal sonograms who did not have surgery.

In summary, despite improvements in sono-graphic technology, detection of small gallstonesremains difficult in some cases. We found thatadherent gallstones can mimic polyps and recom-mend that this appearance be included in thespectrum of sonographic findings in gallstones.Repeat scanning in cases of negative sonogramsbut high clinical suspicion may help decrease thefalse-negative rate for sonographic diagnosis ofgallstones. Finally, reviewing sonograms care-fully for technical adequacy may also reduce theoccurrence of false-negative studies.

REFERENCES

1. Cooperberg PL, Gibney RG. Imaging of the gall-bladder. Radiology 1987;163:605.

2. Marton KI, Doubilet P. How to study the gallblad-der. Ann Intern Med 1988;109:752.

3. Weltman DI, Zeman RK. Acute diseases of the gall-bladder and biliary ducts. Radiol Clin North Am1994;32:933.

4. Shea JA, Berlin JA, Escarce JJ, et al. Revised es-timates of diagnostic test sensitivity and specificityin suspected biliary tract disease. Arch Intern Med1994;154:2573.

5. Hessler PC, Hill DS, Detorie FM, et al. High accu-racy sonographic recognition of gallstones. AJR AmJ Roentgenol 1981;136:517.

6. Hershman MJ, Campion KM, Reilly DT. Can sur-geons rely on ultrasonography for gallstones? J RColl Surg Edin 1986;31:35.

7. Mattson MW, Sterchi JM, Myers RT. Accuracy ofultrasonography and oral cholecystography in thediagnosis of cholelithiasis. Am Surg 1981;47:80.

8. Cooperberg PL, Burhenne HJ. Real-time ultraso-nography diagnostic technique of choice in calcu-lous gallbladder disease. N Engl J Med 1980;302:1277.

9. Boutkan J, Butzelaar MJM, Davies G. The diagno-sis of gallstones—a prospective comparison of oralcholecystography and real-time ultrasound. Neth JSurg 1984;36(5):124.

10. Comerota AJ, Breckenridge J, Maier WP. Surgi-cally proven accuracy of cholecystoechography.Pennsylvania Medicine 1980;83:37.

11. Crade M, Taylor KJ, Rosenfield AT, et al. Surgicaland pathologic correlation of cholecystosonographyand cholecystography. AJR Am J Roentgenol 1978;131:227.

12. Simeone JF, Mueller PR, Ferrucci JT. Significanceof nonshadowing focal opacities at cholecystosonog-raphy. Radiology 1980;137:181.

13. Carroll BA. Gallstones: in vitro comparison ofphysical, radiographic, and ultrasonic characteris-tics. AJR Am J Roentgenol 1978;131:223.

14. Filly RA, Moss AA, Way LW. In vitro investigationof gallstone shadowing with ultrasound tomogra-phy. J Clin Ultrasound 1979;7:255.

15. Grossman M. Cholelithiasis and acoustic shadow-ing. J Clin Ultrasound 1978;6:182.

16. Simeone JR, Ferrucci JT. New trends in gallblad-der imaging. JAMA 1981;246:380.

TABLE 2

False-Negative Rates for Sonographic Detection of Gallstones

FirstAuthor

YearReported

Total No.Patients

No. with Stonesat Surgery

No. Operated onwith NegativeSonograms*

No. with NegativeSonograms and Stones

False-NegativeRate

Camerota10 1980 201 80 10/107 2 2.5%Cooperberg8 1980 993 394 5/586 5 1.3%Mattson2 1981 126 98 19/19 19 19.4%Hessler5 1981 404 69 2/276 1 1.5%Hershmann6 1986 203 191 6/6 6 3.1%Current study 1999 946 934 12/12 12 1.3%

*Out of total with negative sonograms (second number).

MISSED GALLSTONES

VOL. 27, NO. 3, MARCH/APRIL 1999 121