344

Ultrasound-guided

fine-needle biopsy of focal liver lesions: techniques, diagnostic accuracy and complications A retrospective study on 2091 biopsies

Two thousand

and ninety-one

ultrasound-guided

fine-needle

biopsies were performed

in 1946 patients

to diagnose focal

liver lesions. The diagnostic accuracy of fine-needle biopsw is very high (only one false positive was observed), both for aspiration biopsy (93.4%) and for cutting biopsy (95.1%). The difference is not statistically significant. In cases of hepatocellular carcinoma (but not in cases of metastasis or hepatic lymphoma), double biopsy (aspiration and cutting) showed hi&w disenostic sensitivitv than sin& methods. A certain number of benign focal liver lesions were also diagnosed. In the series, no case of death following liver puncture WE observed. It&peritoneal hemanhage was the-most earn-man complication. The risk with a cutting needle being higher than with an aspirative needle.

present

.

-

Fine-needle biopsy (caliber of needle Cl mm) guided by ultrasound (US) or computed tomography (cr) has completely replaced blind percutaneous biopsy for the pa thological diagnosis of focal liver lesions, and has also modified the role of laparoscopy (1,Z). US-guided ftneneedle biopsy (USG-FNB) is safe, inexpensive and has a high diagnostic accuracy (3). Fine-oeedle biopsy (F’NB) can be performed by aspi&ion (using Chihs-type needles) followed by cytological examination, or with cutting needles (usually modified Menghini needles) to obtain tissue fragments for histological examination. Although in most previously published studies of USG-FNB aspirative needles were used with positive results (1,4-L?), there is now a question as to which of the two types of needle has the best diagnostic efficacy for focal liver lesions. Limberg et al. (13) have suggested using cutting needles rather than a+rative

needles to diagnose either primary or secondary liver tumours. Sangatti et al. (14) have suggested that perform@ both aspirative and cuniog needle biopsies together improves fhe diagnosis of hepatoeellular carcinoma (HCC). This is ao important

question as it involves both Safety

and economic problems. lo one study of a MB, for example, four o”t of five pwxturer

397 USGwith fatal

outcome were performed using cutting needles (15). Moreover cutting needles are much more expensive than aspirative needles. The results of a large series of USG-FNB of focal liver le;ians performed in seven Italian centa are reported here. The aims of this study were: (i) to confirm the high diagnostic accuracy of USG-MB in a large series of p+ tients; and (ii) to campwe the diagnostic efficacy of three modalities of USG-F’NB (aspiration biopsy, AB; cutting biopsy, CB; and doub!e biopsy, AB plits CB) for thediag-

“LTIlASOUND-o”mEo

BlOPSY OF LWER

34

LESlONS

n&s of focal liver lesions. Finally we attemy:ed to evaluate complications of these procedures.

(i)Number

MCthOdS

(k) Complications. The data obtained test and? test.

Seven Ilalirn institutions having used USC%MB for at least 3 years participated in this study. Each institution completed the following questionnaire: (a) Beginning time of use of USG-FNB. i.b) Number of hepatic FNB perfomred by 31 December 1988. (e) Number of controtted biopsies. Three fornu of control were accepted: histologic diagnosis (laparaseo~y. surgery and neero~sy), and/or angiographic diagnosis, or diicalUS follow-up for at least 6 months. (d) US-guidance technique (free hand tecbniiue (16): biopsy probe). (e) Type attd caliber of the fine needles. (f) Appttcatioo of rapid staining for the immediate check of aspiration biopsy. (g) Pre- and post-biqxsy monitoring policy. (h) Results of controlted FNB, expressed, w;th regard to the Q”BXlC.2 or the absence Of matignancy, as true positive (TP), true negative (TN), false positive (FF’) negative (FN).

and fake

of insufftcient specimens (i.e., unsuitable for the pztthcicgiist’s interpretation). fj) Distribution of true positive and true negative diagIloses. were elaborated

using Fisher’s exact

The seven institutions had used USG-FNB for an Avery age of 5 years and 3 months (range: 3-8 year+. Two thousand and ninety-one biopsies were performed on 1946 patients, and there were 1920 controlled biopsies. The control was histologic and/or angiogmpbic in 51.5% of the cases(histology in 732 patients; angiogaphy in 148: bistology and angiograpby in 109). In one center the sensitivity of FNB was 78%. while in anotberitwas89.9% and in the atben mo1e than 93%. In six out of sex” centers tile specificity was 100% (Table 1). Seven hundred and sexenty-one biopsies were done wing a free-hand technique (i.e. using an US probe without biopsy apparatus), and 1320 using a biopsy probe (convex, tincar or sector). In each center the FNB were perfamed ty those investigatorx whose names appear as authors of this article. At, the patients had prothrombin activity of more than 50% and a c.xmtofgreaterthan7U WOpermt. No imtitution performed QE”EdieStian. ‘“me centers used local anaesthesia; one center performed biopsies on an in-patient basis only and the others used both in- and out-patients. After the biopsy, aS patients were kept uoder observation for 1-6 h. The needles used for aspimtive biopsy (Chiba needle and spinal needle) had a 23 to 21 Sa”ge fatibe1. ti Ctttthg WdkS (hKUt, HistOC”t) ranged from 22 to 21 Sauges. Two centa qsed immediate stabting Of some cyt&Sic specimens in order to c”al”ate

platelet

the celltdar adequacy of the aspirated material. As a result, in these two bMitutions the average numba of pun* twes for each lesion was 1.3, white in the others it was 2-3. Tbhere were 972 biopsies performed with arpirative

I.. B”SCABml

346 needles and 484 with cutting tteedks. Four hundred and ninety-three biopsies were simultaneouslv performed with both aspirative and cutting neeoies. Tba resttlts from the three bioptic modalities are summarized in Table 2. In one case (aspiration biopsy), FP

et al.

TABLE 4 Distrtbution oftrue wgativs cars

was observed (diagnosis of HCC in the case of regenerative node in cirrhosis). The overall diagnostic accuracy of aspiration vs. cutting biopsy was not significantly different. On the other hand, the diagnostic accuracy of double biopsy was significantly higher. In 29 of 1949 biopsies (1.4%) the specimen was insufficient in the pathologist’s opinion. In Table 3 we compare the results of the three different bioptic modalities in the three groups of patients with different neoplastic diseases. The number of HCC diagnosed by AB, CB or AB plus CB was practically the same. On the other hand two thirds of the metastatic lesions were identified by AB. In the cae of HCC, the sew sitivity of double biopsy was higher than that of single biopsy. No statistical difference was observed in the diagnosis of metastasis. In the case of hepatic involvement of lymphoma. double biopsy showed a higher sensitivity than the single one but the difference was not statistically significant. However, we paint out that the numbers canpared were small. In Table 4 the true negative cases are listed, comparing the diagnosis obteined by FNB and the final diagttosis. The diagnoses of cysts and abscesses were always obtained with aspiration biopsy. In 32 cases of hemangioma a positive diagnosis was reached with biopsy (25 CB and seven AB). In other cases diagnosis was suggested due to the aspiration of blood and confirmed by a follow-up andI

or other imaging techniques.

One hundred and eighty-two

WgafiVe-fOr-maliparq cases were cirrhotic. USG-EB.8 was performed following the US finding of focal echostructural liver alterations. In our investigations there were no deaths. Tbe most important complications (observed in eleven out of 2091 biopsies, that is 0.5%) were: six cases of hemoperito. neum, one case of intrahepatic hematoma, one case of bemobilia. one case of coloperitoneum; one cae of poeumo. thorax and one case of parletal hematoma. Cutting needles were involved in nine out of eleven cases. Blood transfusions were only necessary in one case of hemoperitooeutn (AB of HCC) and in the ca.w of intrabepatic ltematoma (CB of liver abscw).

The USG-MB of focal liver lesions co&ma. in a large series, its high diagnostic effectiveness. as has been previously reported (Table 5). The sensitivity of the technique appears slightly lower in one of the two caters with a 3 year experience. Only one FP resttlt was obsetved. L TP

m

HCC AU ZM I CB 186 0 AB+CB 243 0 AB vs. AB 4 CB:p < 0.0, CB “I. AB + CB: p < 0.01 Asrr. CB: N.S.

M

TOT

24 13

232

6

:z

SENS (0) 89.6 93.4 97.5

every field of malignant pathology examined the diagns tic accuracy of aspiration biopsy (cytological exambratioo) is similar to that of cutting needle biopsy (histologicalexamination). In the case of HCC, but oat met&a& or lymphoma, the use of double biopsy (aspirative and cutting needles) allows increased diagnostic sensitivity. as recently described (14). Rondo et al. (17) have pointed out that the cytologic diagnosis of moderately or poorly differentiated HCC (d&ted as classical HCC) is relatively easy. However, the identification of welldifferentiated HCC from borderline lesions OT large regenerative nodes is very difficult. Oo the other band, histologic criteria (nuclear crowding; microatiar formation; increased cytopksmic basophdia) may allow diagnosis of welldifferentiated HCCeven in the small tissue specimensobtained with fioe cutdog needles. The problem is clinically sigoiticaot. In the repated series, there were 14 coxes of well-d&rentinted HCC and 94 classical cases. However. comparison of cytological samples drawn born tumoral and ooo-tumoral &s hallows the identification of well-differentiated formsof HCC(lO\. Pediobtal. (IS~wiotouttbediaeoos tic effectiveness of identifying hepatocytic naked nuclei io

,

6oe needle aspirates

.

from the liver. Those naked tucki

are ttttmerous in a high percentage of cases of HCC, independad of tumour cytologic typo. This fiodiig is cbamcteristically absent in liver cirrhosis. Finally, a cytologicsl sampk perfomted during kparoscopy yreseots a bigb dianoostic accwacv for HCC (19.20). Furthermore. it &t be emphasizeh that HCC L apa;hology with a high risk of bkediog. In the prosem series, we had one case of bkeding which required blood transfusion. Four cases are reported in a series of Bret et al. (IO), with one death (aspiration biopsy). In a receot review, Fomti et al. (21) stated that in 10 766 cases of abdominal FNB, two deaths occurred due to hemorrhage followiog puncture of HCC. In both cases, cutting needles were involved. In the series of We& et al. (15). three out of fnre deaths were due to the puncture of liver tttmours (two HCC and one secondary mmour) and performed by cutting needles. In cases of wtastasis, the cytological ami histological restdts are simikr. In cases of oletastasfs of lmkoovm origio, cytology can be refitted using immtmocytochemical techniques (22) or replaced by histology (either using fine or coarse needles) to establish the tutuour lineage, neces-

sary when therapeutic decisions are involved. lo the case of lymphoma, our data show that cytological and histological investigations are equally eirective. In these cases cytological studies xo effident when lymphoma has already been determined, otherwise the simultaneous use of immunocytocbemistty or histology is necessary (23,24). In this series USC-MB also correctty diagnosed benign focal liver lesions. However, cysts appear as an indication only in cases with an atypical US picture an&or in an oncc!ogical cootext. The abscess is, on the cootrary. ao eke tive indiation for guided puncture, which, at the same time, can be both a diagnostic and therapeutic step. In focal steatosis, the identification rate of USG-PNB appears high, as recently desaibed (25). Definitive diagoosis of hemangioma was achieved by FNB ia 55% of the c&es without any complication. Recently att identification rate of 100% bas been found usiog USG20 gauge cutting needle biopsy (26). However, nowittvasive tahniqttes (red blood cell ximigmphy (27); dynamic CT (28); MRI (29)) cat, allow correct diagoo& of atypical cases of hemaogioma. The data collected eon&m the low risk of USG-FNB. Hemorrhage was the most frequent complication, but only in two cases were blood transfusions ngessary. We paint out that perattatteous blind biopsies with axuse needles (TN-cut or Meoghioi IA-l.6 mm diameter) of primary or metastatic liver malignancies were followed by fatal complications in 1.7% (three deaths out of 1755 biopsies) (30). Previous and present data suggest that the risk with cutting needle seems to be bighor than that using aspirative needle. Moreover. cuttiog biopsy (and obviously double biopsy) is wxe expensive tltao aspiration biopsy due to the cost of the needle and bistolc+al techniques. fn conclusion, we cootinn the safety and bi& diamtic accuracy of USGFNR. Aspiration biopsy cast be coosidered the first-step technique, which in most cares is diagnostic. The use of immediate staining rededuces the tttunber of p”nctures, without mod5ying diagoostic acouracy as previously published (31). In the case of questionable negativity of AB. owing needle biopsy is advisable. In partlctdar this policy isrequired incke3ofHCC. In this fkld. the routine application of doubk bkpsy can be wefld.

L. BUSCARIN,

et

a,.

the Ever. SigniRcan~c of hsQa,oeytic naked ruclei in the diag”o. aisofbep,~,lularcarcinoma. *cm Cyio, 19RB:32: 437-42. 19 leffcn L, Spisgbsts” 0. Reddy R, e, al. LaQaro~opiEPlly diretied fine “ccdle WQiMb” far the diag”otiir of hspa,ooe”“,ar ca,ci”nna: a safe and BCNIB,Etechnique. Endoscopy ,988; 34: 235-7. 20 CUM X. Marti-Vice”,e A, Mo”esXol1, Wardell F. LaQarormpiccyialogy - A” cvaluaria”. Endoscopy L988.20:ICI-3. 21 Fomari F, Ciwdi 0, Cavanna L, e, 21.CO”IQk3tiO”sof d,,aw. “icaSy g”ided S”c-“cc& abdaninr, b+:opsy.Rsrubs of a multice”,er Mia” sfudv snd review oftbe ,i,cra,we. Seand 1 Gwm-

guided fine needle alpir2,ti.m b&y in e”d”sti”g the-s&iii sanographic abnormaliries. Gac,hx”,e,ology 1987;93: 715-g. 9 Buscarini L. SbO”i G, Cavanna L, * a,. ninirr, and diammstic

GeneI Et, Kanncr M, LokckH, Huh” D. Ti,s valueof b”“,“~oeytodlC”,i~ W”i”g Of ,PQ “& WQiMs i” diag~.+ tic CyMogy. Br I Haemato, ,988: 10: x37-16. 25 Ca,ureSi E, Rapac.zi”i GL, De Simwe F, e, a,. Ul,,w”ogi-&y and echo-guided ,i”c-“ccd,e biopsy in Lc diag,&s of foepl fa,,y liver change. HcpaLogsrUa”,emlogy ,981; 34: 137-40. 26 Cro”a” I,, ErQzza AR, Lhxfman OS, Ridlen MS, Paolclla LP. Ca”emo”a hsma”gkw,m of ,bc Ever: mle of percu,a”co,,s w-y. Radblogy ,988; L66:135-g. 2, Eoge, MA, Markr DS, Sandier MA. Sk,,y Q. Dif,cre”,imio” of fLcal inlrakparic lesions With w”rcrcd blood ccl, imagiq. Raaiology ,983; 146:m-82. 28 A&Ida C. Firhma” EK, Ze,hami E, Herlong F. siegelma” SS. Com~Wd tomography of kpatic cave,“ow kmar.+a. J GmlQ”, Assist TMaOgr ,987; 11:455-60. 7.9 Gb,O”m K, ,,a, Y! Y”sh,kaw* K. 01a,. “epa,ic Nmola: dylllmic MRimagiag. RwJlo!+gy,987: 16.327-31. NI Picdnina F, Sag”a E, PSsqualS G, e, a,. ComQScadool fo,,owisg pcrc”,s”eoua biopsy. A m*ice”,re n,msp3ctiw ahldy 0” ML216biopsies. I Hepat 1986;2: X6-73. 31 Civardi G. FornariF. Cavanna L, DiS,asi M, Sbolli G,Btini L. Vnlw of rapid rrsining and ssessme”, of uhrawurd guided Sne needle aspiration bioprier. Ac,a CybAogka 19SS:32: 552-4. 24 t,ertclJ,