ORIGINAL ARTICLE: Clinical Endoscopy

The role of K-ras gene mutation analysis in EUS-guided FNA cytology specimens for the differential diagnosis of pancreatic solid masses: a meta-analysis of prospective studies Lorenzo Fuccio, MD,1 Cesare Hassan, MD,2 Liboria Laterza, MD,1 Loredana Correale, PhD,3 Nico Pagano, MD,1 Paolo Bocus, MD,4 Carlo Fabbri, MD,5 Antonella Maimone, MD,5 Vincenzo Cennamo, MD,5 Alessandro Repici, MD,6 Guido Costamagna, MD,2 Franco Bazzoli, MD,1 Alberto Larghi, MD, PhD2 Bologna, Italy
Background: Differential diagnosis of pancreatic solid masses with EUS-guided FNA (EUS-FNA) is still challenging in about 15% of cases. Mutation of the K-ras gene is present in over 75% of pancreatic adenocarcinomas (PADC). Objective: To assess the accuracy of K-ras gene mutation analysis for diagnosing PADC. Design: We systematically searched the electronic databases for relevant studies published. Data from selected studies underwent meta-analysis by use of a bivariate model providing a pooled value for sensitivity, speci?city, diagnostic odds ratio, and summary receiver operating characteristic curve. Setting: Meta-analysis of 8 prospective studies. Patients: Total of 931 patients undergoing EUS-FNA for diagnosis of pancreatic solid masses. Intervention: K-ras mutation analysis. Main Outcome Measurements: Diagnostic accuracy of K-ras mutation analysis and of combined diagnostic strategy by using EUS-FNA and K-ras mutation analysis in the diagnosis of PADC. Results: The pooled sensitivity of EUS-FNA for the differential diagnosis of PADC was 80.6%, and the speci?city was 97%. Estimated sensitivity and speci?city were 76.8% and 93.3% for K-ras gene analysis, respectively, and 88.7% and 92% for combined EUS-FNA plus K-ras mutation analysis. Overall, K-ras mutation testing applied to cases that were inconclusive by EUS-FNA reduced the false-negative rate by 55.6%, with a false-positive rate of 10.7%. Not repeating EUS-FNA in cases in which mutation testing of the K-ras gene is inconclusive would reduce the repeat-biopsy rate from 12.5% to 6.8%. Limitations: Small number of studies and between-study heterogeneity. Conclusion: K-ras mutation analysis can be useful in the diagnostic work-up of pancreatic masses, in particular when tissue obtained by EUS-FNA is insuf?cient, and the diagnosis inconclusive. (Gastrointest Endosc 2013;78:596-608.)
Abbreviations: PADC, pancreatic adenocarcinoma; QUADAS, Quality Assessment of Diagnostic Accuracy Studies. DISCLOSURE: All authors disclosed no financial relationships relevant to this publication.
See CME section; p. 637. Copyright ª 2013 by the American Society for Gastrointestinal Endoscopy 0016-5107/$36.00 http://dx.doi.org/10.1016/j.gie.2013.04.162
Received January 10, 2013. Accepted April 3, 2013.
Current affiliations: Department of Medical and Surgical Sciences, S. Orsola-Malpighi Hospital, University of Bologna, Bologna (1), Digestive Endoscopy Unit, Department of Surgery, Catholic University, Rome (2), Im3D Medical Imaging Lab, Turin (3), Endosonography Unit, Veneto Institute of Oncology, Padova (4), Unit of Gastroenterology and Digestive Endoscopy, AUSL Bologna Bellaria-Maggiore Hospital, Bologna (5), Department of Gastroenterology, IRCCS Istituto Clinico Humanitas, Rozzano, Milan, Italy (6).
Reprint requests: Lorenzo Fuccio, Department of Medical and Surgical Sciences, S. Orsola-Malpighi Hospital, University of Bologna, Via Massarenti 9, 40136, Bologna, Italy.
596 GASTROINTESTINAL ENDOSCOPY Volume 78, No. 4 : 2013 www.giejournal.org
Pancreatic cancer represents a signi?cant health problem worldwide, with 1 of the lowest 5-year survival rates of all cancers.1 In 2013, about 46,000 new cases of pancreatic cancer were estimated in the United States,2,3 and the average survival time after diagnosis was 6 months.4 Despite advances in diagnostic imaging techniques and the pivotal role played by EUS and tissue sampling by FNA, the differential diagnosis between malignant and nonmalignant pancreatic masses is still inconclusive in about 15% of cases. A recent meta-analysis demonstrated that EUS-guided FNA (EUS-FNA) with cytology analysis is a highly accurate diagnostic test for solid pancreatic masses, with a pooled sensitivity of 85% and a speci?city of 98%.5 However, the diagnostic accuracy of EUS-FNA is in?uenced by several factors, such as the quantity and quality of material obtained, the size and location of the mass, and the technical skill of the endoscopist as well as the presence of a cytopathologist on site.5-8 Cytopathologic assessment may be dif?cult or not feasible because the material aspirated may be bloody, with scarce or inadequate material. Several methods, mainly based on genetic analyses, have been investigated for improving pancreatic cancer diagnosis.9-11 Pancreatic adenocarcinoma (PADC) has a high incidence of K-ras gene mutations, generally reported in more than 75% of cases,12-14 and these mutations appear to occur early during carcinogenesis.15 For these reasons, K-ras gene mutation analysis has been considered a possible marker for PADC detection. In the last decade, several prospective case series studying detection of K-ras gene mutations in EUS-FNA have been published.16-18 However, the potential impact of K-ras mutation determination on the accuracy of EUS-FNA for PADC and its role in the diagnostic algorithm for pancreatic masses are still unclear. The aim of our study was to perform a structured metaanalysis of the available evidence on the diagnostic accuracy of K-ras gene mutation detection in pancreatic solid mass lesions.
METHODS
Data sources and searches A protocol was written before the meta-analysis was carried out. We identi?ed relevant studies by searching PubMed, EMBASE, Google Scholar, Scopus, and the Cochrane Library. We searched the literature without language restriction through December 31, 2012. Search terms were K-ras or Kras and endoscopic ultrasound or EUS or ?ne-needle aspiration or FNA and pancreas or cancer. In addition, we identi?ed relevant studies from the reference list of each selected article. We also handsearchedabstractspresented through 2012at theAmerican Gastroenterological Association Digestive Disease Week,
Take-home message
EUS-guided FNA (EUS-FNA) plus K-ras testing in cases of inconclusive results increases the overall sensitivity for diagnosis of pancreatic masses.
K-ras gene mutation analysis reduces the need for repeating EUS-FNA in cases in which EUS-FNA results are inconclusive.
United European Gastroenterology Week, and at the Italian National Congress of Gastroenterology. Selection criteria were established a priori to minimize bias.19 Inclusion and exclusion criteria are summarized in Table 1. When we found multiple articles for a single study, we used the latest publication and supplemented it, if necessary, with data from the previous publications. If any clari?cation of data was necessary, we contacted the authors for detailed information. Eligibility assessment was performed independently by 2 reviewers (L.F., L.L.).
Data extraction and quality assessment Three investigators independently extracted data on the following items from the selected studies: year of publication, location of the study, number of centers involved, type of publication (full-text or abstract form), enrollment period, number of patients enrolled, sex, size of the diagnostic FNA needle (ie, 19, 22, or 25 gauge), number of needle passes (mean and/or range), the presence of an on-site cytopathologist, number of PADCs according to the ?nal diagnosis, number of non-PADC lesions, number of PADC and non-PADC lesions diagnosed by EUS-FNA, number of inconclusive results on EUS-FNA (inconclusive de?ned as insuf?cient material, atypia, or suspicion of malignancy), number of inconclusive results on EUS-FNA in PADC and non-PADC groups, type of DNA analysis for K-ras gene mutation detection, number of codons analyzed (codons 12, 13, and 61), number of cases in which K-ras mutation analysis was successful, number of PADC and non-PADC cases in which K-ras was mutated, number of cases with inconclusive EUS-FNA results (total and according to PADC and non-PADC groups) in which K-ras was mutated. The Quality Assessment of Diagnostic Accuracy Studies (QUADAS) questionnaire was used to assess the quality of the selected studies.20 Items were rated as yes, no, or unclear. Disagreements were resolved by discussion.
Statistical analysis For each study, a 2  2 contingency table was constructed that compared the ?nal disease diagnosis with test results. The ?nal diagnosis was established by cytologic and/or histologic examinations, the histopathologic examination of the surgically resected specimen, and the results of other diagnostic investigations or clinical follow-up. For the purpose of this meta-analysis, cases that were
Fuccio et al EUS-FNA and K-ras testing for pancreatic solid mass diagnosis
www.giejournal.org Volume 78, No. 4 : 2013 GASTROINTESTINAL ENDOSCOPY 597
inconclusive by EUS-FNA were considered as negative results, being classi?ed as either false negative or true negative according to the ?nal diagnosis. Descriptive statistics for the dataset included sensitivity, speci?city, and false-positive rate of the primary studies, their positive and negative likelihood ratios, and their diagnostic odds ratios (OR). The degree of variability among study results was ?rst evaluated graphically by plotting sensitivity and speci?city from each study on a forest plot. The chi-square test was performed to assess heterogeneity of studies results, the null hypothesis being in both cases that all are equal. The I2 statistic provides an estimate of the amount of variance due to heterogeneity rather than chance and is based on the traditional measure of variance,
the Cochrane Q statistic. Values of I2 equal to 25%, 50%, and 75% were assumed to represent low, moderate, and high heterogeneity, respectively. We used a bivariate model for diagnostic meta-analysis to obtain an overall sensitivity and an overall speci?city.21 The bivariate model uses a random-effects approach for both sensitivity and speci?city, which allows for betweenstudy variability. To graphically present the results, we plotted the individual and summary points of sensitivity and speci?city in a receiver operating characteristic graph, plotting the index test’s sensitivity (true-positive rate) on the y axis against 1-speci?city (false-negative rate) on the x axis. In addition, we plotted a 95% prediction region around the pooled estimates to illustrate the precision
TABLE 1. Inclusion and exclusion criteria
Inclusion criteria Exclusion criteria Prospective study Retrospective study
Used EUS-guided FNA for diagnosis of solid pancreatic masses Used abdominal US-guided FNA
K-ras mutation analysis based on the material obtained by FNA Not pancreas specific
Used a reference standard of definitive surgical histology, unequivocal histo-cytopathology, or clinical follow-up
Review articles, case reports, editorials, and corresponding letters that did not report their own data
Data available to construct contingency tables for true-positive, false-positive, false-negative, and true-negative determinations
Insufficient data (not completed even after directly contacting first and/or corresponding authors)
Figure 1. Flowchart demonstrating the algorithm for identifying suitable articles for inclusion. AGA-DDW, American Gastroenterological AssociationDigestive Disease Week; UEGW, United European Gastroenterology Week; FISMAD, Italian National Congress of Gastroenterology.
598 GASTROINTESTINAL ENDOSCOPY Volume 78, No. 4 : 2013 www.giejournal.org
EUS-FNA and K-ras testing for pancreatic solid mass diagnosis Fuccio et al
with which the pooled values were estimated (con?dence ellipse of a mean) and to show the amount of betweenstudy variation (prediction ellipse; the likely range of values for a new study). Diagnostic accuracies of the different tests were compared by using a logistic mixed-effect model in which the primary studies were considered as random effects, with test as ?xed effects. These analyses were undertaken by using R statistical software (R package version 0.5.1.mada: Meta-analysis of Diagnostic Accuracy (mada); R Core Team (2012). R Foundation for Statistical Computing, Vienna, Austria.).
RESULTS Eligible studies and quality assessment As shown in Figure 1, the literature search identi?ed 8 studies published (7 in extensus10,16-18,22-24 and 1 in abstract form25) that ful?lled the inclusion criteria. Their main characteristics are reported in Table 2. Overall, 931 cases were entered in the meta-analysis, ranging from 34 to 394 patients per study. EUS-FNA was technically successful in all patients, almost all studies used the same size FNA needle (22 gauge), and the same mean number of needle passes per
patient was performed across all the studies. Notably, K-ras gene mutation analysis was feasible in all cases independently of the adequacy of the cellularity obtained by FNA. Details on EUS-FNA and K-ras gene mutation analyses are reported in Table 3. The quality of the eligible studies, as assessed according to the QUADAS criteria, is reported in Figure 2. The percentage of high-quality studies (ie, those for which a yes response applied) ranged from 66% to 75% for each of the 12 items. In most of the studies, it was unclear whether the patients received the same reference standard regardless of the index test result and whether the authors interpreted the reference standard results without knowledge of the results of the K-ras gene mutation analysis.
Synthesis of results The results of the included individual studies are provided in Table 4.
EUS-FNA In 116 cases (12%), the EUS-FNA material was deemed as inconclusive. Of these, 88 cases (76%) were false negatives (ie, PADC at the ?nal diagnosis), and 28 were true
TABLE 2. Study characteristics of included studies
Study (reference) Year Country
No. of centers
Enrollment period Patients
Final diagnosis*
PADC Total
Non-PADC total lesions
Non-PADC benign lesions
Other pancreatic neoplasia Tada (22) 2002 Japan 1 Feb 1998-Mar 2001 34 1,2,3,4,5 26 8 8 (CP) 0
Pellisé (18) 2003 Spain 1 Sep 2001-Mar 2002 57 1,5 33 24 5 (CP) 19 (6 PNET, 5 IPMN, 5 CyA, 2 ME, 1 Ly)
Takahashi (23) 2005 Japan 1 Aug 1998-Apr 2003 77 1,5 62 15 15 (CP) 0
Maluf-Filho (17) 2007 Brazil 1 May 2002-Apr 2004 74 1,5 57 17 11 (CP) 6 (PNET)
Bournet (16) 2009 France 4 Jan 2005-Apr 2007 178 1,2,5 129 49 33 (27 CP, 6 BI)
16 (12 PNET, 4 ME)
Wang (10) 2011 China 1 Jan 2008-Mar 2010 82 1,2,3,5 54 28 17 (10 CP, 3 AIP, 1 BI, 1 PT, 2 PP)
11 (4 IPMN, 7 MCN)
Ogura (24) 2012 Japan 1 Mar 2004-Sep 2009 394 1,2,5 307 87 47 (24 CP, 23 AIP)
40 (20 PNET, 8 ME, 3 ACC, 3 Ly, 3 SPT, 2 SCT, 1 LC)
Visani (25) 2012 Italy 1 Nov 2010-Oct 2011 35 1,2,3,5 18 17 7 (5 PP; 2 SCT)
10 (4 PNET, 1 ME, 2 IPMN, 1 MCN, 2 SPT)
PADC, Pancreatic adenocarcinoma; CP, chronic pancreatitis; PNET, pancreatic neuroendocrine tumor; IPMN, intraductal papillary mucinous tumor; CyA, cystadenoma; ME, metastasis; Ly, lymphoma; BI, benign inflammation sequelae of inflammatory pancreatitis; AIP, autoimmune pancreatitis; PT, pancreatic tuberculosis; PP, pancreatic pseudocyst; IPMN, intraductal papillary mucinous tumor; MCN, mucinous cystic neoplasm; ACC, acinar cell carcinoma; SPT, solid-pseudopapillary tumor; SCT, serous cystic tumor; LC, lymphoepithelial cyst. *Final diagnosis: 1 Z surgery; 2 Z histologic/cytologic examinations; 3 Z imaging techniques; 4 Z autopsy; 5 Z follow-up.
www.giejournal.org Volume 78, No. 4 : 2013 GASTROINTESTINAL ENDOSCOPY 599
Fuccio et al EUS-FNA and K-ras testing for pancreatic solid mass diagnosis
negatives (ie, negative at the ?nal diagnosis). The sensitivity for PADC detection ranged from 61% to 93%, and speci?city ranged from 92% to 99% (Fig. 3). Estimated sensitivity and speci?city was 80.6% (95% con?dence interval [CI], 72.1-86.9) and 97.0% (95% CI, 93-99%). Between-study heterogeneity was substantial, with an I2 of 76.5% for sensitivity (P ! .001). The reverse was true
for speci?city, with an I2 of 0% (P Z .588). The summary receiver operating characteristic curve (SROC) graph for the diagnosis of PDAC is shown in Figure 4. The partial area under the curve (restricted to observed falsepositive rates) was 73.7%. A positive correlation across studies was detected between sensitivity and speci?city, not the negative correlation that would be expected.
TABLE 3. Details on EUS-FNA (technically successful in all cases) and K-ras gene mutation analysis
Study (reference)
EUS-FNA sample adequate (%)
Size of needle, gauge
Mean no. needle passes/patient
On-site cytopathologist
K-ras gene mutation analysis Codon
K-ras gene mutation analysis successful, % Tada (22) 71 22 2.5 No Mutation specific 12 100
Pellisé (18) 93 22 2.4 Yes PCR 12 100
Takahashi (23) 92 22 2.3 Yes PCR 12 100
Maluf-Filho (17) 93 22 3 Yes PCR 12 100 Bournet (16) 84 22 At least 2/patient No PCR þ sequencing 12, 13 100 Wang (10) 80 19, 22 2.6 No PCR þ sequencing 12, 13 100 Ogura (24) 90 22 2.3 Yes Mutation specific 12 99.7 Visani (25) 80 19, 22, 25 2.4 Yes Mutation specific þ sequencing 12, 13, 61 100
EUS-FNA, EUS-guided FNA; PCR, polymerase chain reaction.
Figure 2. The quality of the eligible studies as assessed according to the 12 items included in the Quality Assessment of Diagnostic Accuracy Studies20 (QUADAS) criteria.
600 GASTROINTESTINAL ENDOSCOPY Volume 78, No. 4 : 2013 www.giejournal.org
EUS-FNA and K-ras testing for pancreatic solid mass diagnosis Fuccio et al
TABLE 4. Results of individual studies for EUS-FNA and K-ras gene mutation analysis
Study (reference)
EUS-FNA K-ras EUS-FNA D K-ras TP FP FN TN TP FP FN TN TP FP FN TN Tada (22) 16 0 10 8 20 0 6 8 21 0 5 8
Pellisé (18) 31 0 2 24 24 0 9 24 32 0 1 24
Takahashi (23) 52 0 10 15 46 0 16 15 58 0 4 15
Maluf-Filho (17) 47 1 10 16 38 1 19 16 48 2 9 15
Bournet (16) 108 0 21 49 79 1 50 48 115 1 14 48
Wang (10) 33 0 21 28 48 9 6 19 43 9 11 19
Ogura (24) 268 0 39 87 267 3 40 84 286 3 21 84
Visani (25) 16 0 2 17 15 1 3 16 18 1 0 16
EUS-FNA, EUS-guided FNA; TP, true positive; FP, false positive; FN, false negative; TN, true negative.
Figure 3. EUS-guided FNA. Forest plots show the sensitivity and speci?city with 95% CIs for each individual study. Estimated sensitivity and speci?city was 80.6% (95% CI, 72.1%-86.9%) and 97.0% (95% CI, 93%-99%). In this graph, between-study variability is also provided, showing substantial heterogeneity for sensitivity. CI, con?dence interval.
www.giejournal.org Volume 78, No. 4 : 2013 GASTROINTESTINAL ENDOSCOPY 601
Fuccio et al EUS-FNA and K-ras testing for pancreatic solid mass diagnosis
K-ras K-ras gene mutation analysis was feasible in all 931 patients. The pooled sensitivity (based on a bivariate approach) was 76.8% (95% CI, 68-84), and the overall speci?city was 93.3% (95% CI, 84.9%-97.2%). As shown in Figure 5, there was substantial heterogeneity for both sensitivity (I2 84.1%; P ! .001) and speci?city (I2 Z73.7%; P! .001). Figure 6 plots the SROC with the summary operating point for PADC diagnosis. The partial area under the curve was 79.9%.
Combination of EUS-FNA with K-ras mutation determination The pooled sensitivity and speci?city for PADC diagnosis, calculated by using data from all studies, was 88.7% (95% CI, 83.6%-92.4%) and 92.0% (95% CI, 83.0%-96.5%). There was signi?cant heterogeneity across studies for both sensitivity (I2 Z 75.1%; P Z .007) and speci?city (I2 Z 64.6%; P! .001) (Fig. 7). Figure 8 plots the SROC curve with the summary operating point for PADC diagnosis. The partial area under the curve was 83.9%.
Comparison of SROC curves K-ras mutation determination versus EUS-FNA plus K-ras mutation determination. The summary estimates of sensitivity and speci?city are well-separated (Fig. 9). Furthermore, logistic regression analysis also indicated that EUS-FNA in combination with K-ras mutation analysis offered a better accuracy for PADC diagnosis (P ! .001) than K-ras gene analysis alone. It would be
safe to conclude that EUS-FNA in combination with K-ras mutation determination is a more reliable way for PADC diagnosis than K-ras mutation determination alone. EUS-FNA versus EUS-FNA plus K-ras mutation determination. The summary estimates of sensitivity and speci?city are well-separated (Fig. 10). Logistic regression analysis also indicated that EUS-FNA in combination with K-ras analysis offered a better accuracy for PADC diagnosis than EUS-FNA determination alone (P Z .008). It would be safe to conclude that EUS-FNA in combination with K-ras mutation determination is a more reliable way for PADC diagnosis than EUS-FNA alone.
K-ras mutation determination in inconclusive EUS-FNA results Of the 116 patients with inconclusive EUS-FNA results, 88 patients (75.9%) were eventually diagnosed with PADC. Overall, K-ras was positive in 52 of the inconclusive cases (45%), correctly classifying 49 of the 88 PADCs, corresponding to a pooled sensitivity and speci?city of 56% and 89%, respectively. The overall absolute number of PADCs diagnosed when we passed from EUS-FNA to a sequential EUS-FNA þ K-ras strategy increased from 572 of 931 (61%) to 604 of 931 (65%).
DISCUSSION
According to our meta-analysis, K-ras gene mutation analysis is an accurate technique for diagnosing PADC in patients with pancreatic solid masses, with an overall sensitivity and speci?city of 76.8% and 93.3%, respectively. Importantly, our meta-analysis showed a potential synergism between K-ras testing and EUS-FNA, with a major role of K-ras mutation determination for those cases in which EUS-FNA was inconclusive. In fact, when we applied K-ras mutation determination to cases in which EUS-FNA was inconclusive, the false-negative rate was reduced by 55.6%, with a false-positive rate of 10.7%. However, it should be pointed out that the combination strategy EUS-FNAþK-ras testing, although it increases the sensitivity by 8% when compared with the EUS-FNA–only approach, decreases the speci?city by 5%. Despite the substantial reduction of the number of false-negative cases, the combined technique represents an undeniable advantage, and the risk, albeit small, of false-positive results prompts a cautious integration of the combined strategy results with all the other clinical variables of the patient. The K-ras gene is the most commonly mutated oncogene in pancreatic cancers (O75% of cases), generally by point mutations in codon 12.12-14,26,27 The K-ras gene is located on chromosome 12p and encodes a membrane-bound guanosine triphosphate (GTP)–binding protein, which mediates various cellular functions, such as proliferation and cellular survival; once mutated, the regulated GTPase activity is abolished, which results in
Figure 4. EUS-guided FNA. In this ?gure, the SROC curve is plotted. The point estimate of the pair of sensitivity (80.6%; 95% CI, 72.9%-86.9%) and false-positive rate (0.029; 95% CI, 0.012-0.071) also is plotted. CI, con?dence interval; SROC, summary receiver operating characteristic; EUSFNA, EUS-guided FNA.
602 GASTROINTESTINAL ENDOSCOPY Volume 78, No. 4 : 2013 www.giejournal.org
EUS-FNA and K-ras testing for pancreatic solid mass diagnosis Fuccio et al
constitutive signalling.28 Mutations in the K-ras gene are considered 1 of the earliest genetic events in pancreatic tumorigenesis.15,29,30 Furthermore, a stepwise increase in K-ras mutations, with an increasing grade of dysplasia in pancreatic intraepithelial neoplasia from patients with ductal adenocarcinoma has been found.15 The management of patients with pancreatic solid masses and inconclusive FNA results has not been widely standardized. According to the European Society of Gastrointestinal Endoscopy clinical guidelines, in patients with inconclusive ?ndings at initial EUS-FNA, repetition of EUSFNA is strongly advised.31 However, published studies on this issue have shown that the repetition of EUS-FNA may have suboptimal accuracy, yielding a correct diagnosis in about 60% to 80% of cases.32-34 In the retrospective study of Nicaud et al,32 based on 28 patients, repeating EUS-FNA inpatientswithpancreaticsolidmassesprovidedasensitivity for the diagnosis of cancer of 35% and an overall accuracy of 61%. In a similar study including 24 patients, Eloubeidi et al33 found a diagnostic accuracy of 84%. It should be highlighted that both studies were performed in tertiary
care referral centers by highly experienced endosonographers, thus limiting their external validity. Based on the results of our meta-analysis, the implementation of K-ras gene mutation analysis for cases in which EUS-FNA was inconclusive would reduce the need for repeat EUS-FNA. It could be argued that K-ras sensitivity appears to be reduced substantially when K-ras sensitivity for PADC is assessed only in cases in which EUS-FNA results are inconclusive. This decrease in the sensitivity may in part be explained by the heterogeneous distribution of the K-ras gene mutation within the tumor mass and may re?ect the scarcity of biologic material that usually characterized cases in which the EUS-FNA diagnosis was inconclusive. On the other hand, K-ras mutation analysis was feasible in all cases, independent of the adequacy of cellularity obtained by FNA, suggesting that other factors may be responsible for the decreased sensitivity observed in specimens from cases in which the EUS-FNA diagnosis was inconclusive. The impact of such suboptimal sensitivity is worsened by the extremely high prevalence of PADC in the study
Figure 5. K-ras testing. Forest plots show the sensitivity and speci?city with 95% CIs for each individual study. The pooled (based on a bivariate approach) sensitivity was 76.8% (95% CI, 0.68-0.84), and the overall speci?city was 93.3% (95% CI, 84.9-97.2). In this graph, we can view the results of variation across studies. For both sensitivity and speci?city, there was substantial heterogeneity. CI, con?dence interval.
www.giejournal.org Volume 78, No. 4 : 2013 GASTROINTESTINAL ENDOSCOPY 603
Fuccio et al EUS-FNA and K-ras testing for pancreatic solid mass diagnosis
population and especially in those cases in which the EUSFNA ?ndings were inconclusive, in which a PADC prevalence rate of nearly 80% was estimated by our analysis. Therefore, a possible synergism between EUS-FNA and K-ras gene mutation testing could contribute to limiting the repetition of EUS-FNA only to those cases that are negative at K-ras gene mutation analysis (ie, wild-type K-ras), while considering as true positive cases in which K-ras testing is positive after inconclusive EUS-FNA results. Indeed, not repeating EUS-FNA in those cases with inconclusive results for K-ras gene mutation testing would reduce the repeat-biopsy rate from 12.5% to 6.8%. Avoiding further biopsy would lead to a substantial saving of ?nancial and medical resources. It should be noted that the total cost for K-ras assay, including sequencing, is estimated 60 dollars. Furthermore, K-ras gene mutation analysis is not operator-dependent and is feasible in almost all cases. On the other hand, a repeat EUS-FNA with the presence of an on-site cytopathologist may cost more than 1300 to 1600 dollars, and repeat EUS-FNA has a somewhatcomparableorslightlyhigherdiagnosticaccuracy. Administration of serum CA 19-9 (carbohydrate antigen) is routinely performed in patients with pancreatic solid masses. A recently published meta-analysis, including data from 57 studies with more than 3200 patients with pancreatic cancer and 1800 with benign pancreatic disease, showed that the summary estimates for CA 19-9 (cutoff R37 U/mL) were 78.2% mean sensitivity and 82.8% mean speci?city for discriminating pancreatic carcinoma from benign pancreatic disease.35 Based on the results of our meta-analysis, the only K-ras testing presented
a pooled sensitivity and speci?city of 78.3% and 93.9%, respectively. The study by Wang et al,10 including 82 patients, evaluated the diagnostic accuracy of the combination of both tests, K-ras testing and serum CA 19-9, in the diagnosis of cases in which EUS-FNA results were indeterminate and demonstrated that the sensitivity of the combination strategy was signi?cantly higher than for serum CA 19-9 alone, but with no differences in the other diagnostic parameters. Further studies, with larger sample size, are warranted to con?rm the bene?t of the combination strategy. Our meta-analysis presents some limitations. The meta-analysis was focused on the role of K-ras gene mutation analysis in solid pancreatic masses, therefore any inference to cystic lesions is not appropriate. However, it should be pointed out that 4 of the 8 included studies also included some cystic lesions in the non-PADC group10,18,24,25; in particular, a total of 36 cystic lesions of a total of 245 non-PADC lesions (14.7%) were included, as reported in Table 2. Because of a paucity of data, we were not able to exclude cystic lesions from the analyses. The prevalence of K-ras gene mutations has been reported to range from 0% to 42% in benign cystic lesions and from 20% to 53% in malignant lesions.36-39 The role of K-ras testing in the differential diagnosis between mucinous and nonmucinous, malignant and benign cystic neoplasms is still unclear.36,38,40 In a study of 36 patients, K-ras mutation combined with the loss of heterozygosity had a sensitivity of 91% and a speci?city of 93% for the diagnosis of malignant cysts.40 Subsequently, Khalid et al38 performed a larger study, including 113 patients, and reported that the presence of the K-ras mutation did not differ between malignant and premalignant cysts, and the combined strategy still continued to have a high speci?city (94%) but a substantially decreased sensitivity (37%). Several other studies have con?rmed the low sensitivity and the high speci?city of K-ras mutation testing in the differentiation of benign and malignant pancreatic cysts.41 K-ras mutation analysis in cystic lesions may provide useful information, but published data suggest that it cannot be recommended as the only test but always should be considered in addition to other genetic analyses (ie, loss of heterozygosity) and diagnostic modalities (ie, dosage of cyst ?uid carcinoembryonic antigen). K-ras mutation testing might also play a role as a prognostic factor; indeed, it has been shown recently that the detection of the K-ras mutation is an independent risk factor signi?cantly associated with a non-benign course of cystic lesions (OR 3.4).39 Only prospective studies were included in our metaanalysis, therefore, although an extensive literature search was performed, only 8 studies were included: 7 were published in extensus and 1 in abstract form; in this latter case, the authors were directly contacted and provided further information. Three studies were excluded because of insuf?cient data to construct contingency tables: 2 published in
Figure 6. K-ras testing. In this ?gure, the SROC curve is plotted. The point estimate of the pair of sensitivity (76.6%; 95% CI, 67.9%-83.7%) and false-positive rate (0.067; 95% CI, 0.028-0.151) is also plotted. SROC, summary receiver operating characteristic.
604 GASTROINTESTINAL ENDOSCOPY Volume 78, No. 4 : 2013 www.giejournal.org
EUS-FNA and K-ras testing for pancreatic solid mass diagnosis Fuccio et al
abstract form42,43 and 1 published as full text.11 In 2 cases, e-mails with requests for further details were sent to ?rst and/or corresponding authors11,42; in 1 case we were not able to ?nd a contact address.43 Potential factors of heterogeneity were evaluated. The studies did not differ according to the needle size used and the mean number of needle passes per patient; most of the studies17,18,23-25 included an on-site cytopathologist. All the studies used the currently suggested methods for mutation analysis (polymerase chain reaction, mutation speci?c and direct sequencing analyses),44 however, unavoidably, different methodologies present different diagnostic accuracy.45 The true rate of K-ras mutation in carcinoma of the pancreas has been widely debated during recent years because it is directly in?uenced by several variables, including the methodology implemented.46 Considering the impact that K-ras gene mutation detection has in clinical practice (ie, management of
metastatic colon cancer), it is almost surprising that a widely accepted consensus on the methodology to use for K-ras gene analysis has not yet been proposed. Every mutation detection technique presents drawbacks; sequencing, for example, is a highly speci?c technique withaverylowfalse-positiverate,butitisnotverysensitive; mutationspeci?ctechniquesdependontheoriginaldesign of the assay, testing for a subset of the most common mutations and leading to false negatives when different mutations are present. Several promising detection techniques with high sensitivity and speci?city have been proposed in recent years (ie, peptide nucleic acid–directed polymerase chain reaction), and validating studies are warranted. The implementation in the future of K-ras gene mutation testing assays with increased sensitivity and speci?city might theoretically further increase the usefulness of K-ras mutation testing in cases in which EUSFNA results are indeterminate. The discrepant values for
Figure 7. Combination strategy EUS-FNA plus K-ras testing. Forest plots show the sensitivity and speci?city with 95% CIs for each individual study. The pooled sensitivity and speci?city for PADC diagnosis, calculated using data from all studies, was 88.7% (95% CI, 83.6%-92.4%) and 92.0% (95% CI, 83.0%-96.5%), respectively. In this graph, we can view the results of variation across studies. For both sensitivity and speci?city, there was substantial heterogeneity. EUS-FNA, EUS-guided FNA; PADC, pancreatic adenocarcinoma.
www.giejournal.org Volume 78, No. 4 : 2013 GASTROINTESTINAL ENDOSCOPY 605
Fuccio et al EUS-FNA and K-ras testing for pancreatic solid mass diagnosis
sensitivity and speci?city associated with different assays used in the studies included in our meta-analysis might haveconditionedtheresultsofouranalysisandcontributed to the heterogeneity.
All the included studies in our analysis evaluated mutation at codon 12; 2 studies evaluated both codons 12 and 13,10,16 and only 1 study analyzed 3 codons, 12, 13, and 61.25 It should be pointed out that mutations at codons 13 and 61 are generally rare.47 Indeed, in the study of Bournet et al,16 no mutation of the K-ras gene at codon 13 was detected in codon 12, K-ras–negative samples. In the study of Wang et al,10 of the 57 patients with K-ras mutations, 56 had mutations at codon 12, and only 1 patient with pancreatic cancer had a mutation at codon 13. Similarly, in the study of Visani et al,25 no mutations at codon 13 were detected but only at codon 12 and in 4 cases also at codon 61. Therefore, it is unlikely that the lack of analysis of codons 13 and 61 in some studies might have substantially in?uenced the K-ras gene mutation detection rates. The quality assessment was performed according to the 12-item QUADAS questionnaire20 (Fig. 2), and a positive response was applied for most of the items. Therefore, the included studies can be considered high-quality studies. Notably, almost all studies did not mention whether or not the reference standard was independent of the results of the index test (K-ras status), therefore it is not possible to exclude the presence of incorporation bias for some of the included studies. Furthermore, in 4 studies,16,22,24,25 the results of the reference standard were interpreted with knowledge of the K-ras status, therefore a review bias cannot be avoided. Finally, patients without PADC and without benign pancreatic masses differed among the studies (Table 1). Subgroup analysis, however, could not be performed because of a paucity of
Figure 8. Combination strategyEUS-FNAplus K-rastesting. In this ?gure, the SROC curve is plotted. The point estimate of the pair of sensitivity (88.7%; 95% CI, 83.4%-92.4%) and false-positive rate (0.08; 95% CI, 0.035-0.17) is also plotted. EUS-FNA, EUS-guided FNA; SROC, summary receiver operating characteristic.
Figure 9. Comparisonof SROC curves: EUS-FNA plus K-ras testing versus K-ras testing alone. SROC, summary receiver operating characteristic; EUS-FNA, EUS-guided FNA.
Figure 10. Comparison of SROC curves: EUS-FNA plus K-ras testing versus EUS-FNA alone. SROC, summary receiver operating characteristic; EUS-FNA, EUS-guided FNA.
606 GASTROINTESTINAL ENDOSCOPY Volume 78, No. 4 : 2013 www.giejournal.org
EUS-FNA and K-ras testing for pancreatic solid mass diagnosis Fuccio et al
data to construct contingency tables in most of the included studies. In conclusion, K-ras mutation analysis can be useful in the diagnostic work-up of pancreatic mass lesions and may complement other diagnostic modalities, in particular whenEUS-FNAcytologyspecimensarejudgedinconclusive. Inthesecases,discoveryofaK-rasgenemutationcanspare an unnecessary repeat EUS-FNA procedure. An additional FNA pass during EUS for solid pancreatic masses could be performed and used for K-ras mutation testing in case of an inconclusive diagnosis (Fig. 11). However, when adopting K-ras mutation analysis as an additional test in the diagnosis of solid pancreatic masses, the clinician should be aware that the substantial reduction in the falsenegative rate is counterbalanced by a relatively small increaseinthefalse-positiverate.Therefore,K-rasmutation testing should alwaysdas any other similar modalitydbe cautiously interpreted within the clinical context. Further studies are needed to con?rm this ?nding.
REFERENCES
1. Warshaw AL, Fernandez-del Castillo C. Pancreatic carcinoma. N Engl J Med 1992;326:455-65. 2. Siegel R, Desantis C, Virgo K, et al. Cancer treatment and survivorship statistics, 2012. CA Cancer J Clin 2012;62:220-41. 3. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2013. CA Cancer J Clin 2013;63:11-30. 4. Bornman PC, Beckingham I. ABC of liver, pancreas and gallbladder. Pancreatic tumours. BMJ 2001;322:721-3. 5. Hewitt MJ, McPhail MJ, Possamai L, et al. EUS-guided FNA for diagnosis of solid pancreatic neoplasms: a meta-analysis. Gastrointest Endosc 2012;75:319-31. 6. Harewood GC, Wiersema MJ. Endosonography-guided fine needle aspiration biopsy in the evaluation of pancreatic masses. Am J Gastroenterol 2002;97:1386-91.
7. Buscail L, Faure P, Bournet B, et al. Interventional endoscopic ultrasound in pancreatic diseases. Pancreatology 2006;6:7-16. 8. Chang KJ, Nguyen P, Erickson RA, et al. The clinical utility of endoscopic ultrasound-guided fine-needle aspiration in the diagnosis and staging of pancreatic carcinoma. Gastrointest Endosc 1997;45:387-93. 9. Levy MJ, Oberg TN, Campion MB, et al. Comparison of methods to detect neoplasia in patients undergoing endoscopic ultrasound-guided fine-needle aspiration. Gastroenterology 2012;142:1112-1121 e2. 10. Wang X, Gao J, Ren Y, et al. Detection of KRAS gene mutations in endoscopic ultrasound-guided fine-needle aspiration biopsy for improving pancreatic cancer diagnosis. Am J Gastroenterol 2011;106:2104-11. 11. Salek C, Benesova L, Zavoral M, et al. Evaluation of clinical relevance of examining K-ras, p16 and p53 mutations along with allelic losses at 9p and 18q in EUS-guided fine needle aspiration samples of patients with chronic pancreatitis and pancreatic cancer. World J Gastroenterol 2007;13:3714-20. 12. Koorstra JB, Hustinx SR, Offerhaus GJ, et al. Pancreatic carcinogenesis. Pancreatology 2008;8:110-25. 13. Kimura W, Zhao B, Futakawa N, et al. Significance of K-ras codon 12 point mutation in pancreatic juice in the diagnosis of carcinoma of the pancreas. Hepatogastroenterology 1999;46:532-9. 14. Almoguera C, Shibata D, Forrester K, et al. Most human carcinomas of the exocrine pancreas contain mutant c-K-ras genes. Cell 1988;53: 549-54. 15. Lohr M, Kloppel G, Maisonneuve P, et al. Frequency of K-ras mutations in pancreatic intraductal neoplasias associated with pancreatic ductal adenocarcinoma and chronic pancreatitis: a meta-analysis. Neoplasia 2005;7:17-23. 16. Bournet B, Souque A, Senesse P, et al. Endoscopic ultrasound-guided fine-needle aspiration biopsy coupled with KRAS mutation assay to distinguish pancreatic cancer from pseudotumoral chronic pancreatitis. Endoscopy 2009;41:552-7. 17. Maluf-Filho F, Kumar A, Gerhardt R, et al. Kras mutation analysis of fine needle aspirate under EUS guidance facilitates risk stratification of patients with pancreatic mass. J Clin Gastroenterol 2007;41:906-10. 18. Pellisé M, Castells A, Gines A, et al. Clinical usefulness of KRAS mutational analysis in the diagnosis of pancreatic adenocarcinoma by means of endosonography-guided fine-needle aspiration biopsy. Aliment Pharmacol Ther 2003;17:1299-307. 19. GreenlandS.Meta-analysis.In:RothmanKJ,GreenlandS,editors.Modern Epidemiology. 2nd ed. Philadelphia: Lippincot-Raven; 1998. p. 643-73. 20. Whiting P, Rutjes AW, Reitsma JB, et al. The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews. BMC Med Res Methodol 2003;3:25. 21. Reitsma JB, Glas AS, Rutjes AW, et al. Bivariate analysis of sensitivity and specificity produces informative summary measures in diagnostic reviews. J Clin Epidemiol 2005;58:982-90. 22. Tada M, Komatsu Y, Kawabe T, et al. Quantitative analysis of K-ras gene mutation in pancreatic tissue obtained by endoscopic ultrasonography-guided fine needle aspiration: clinical utility for diagnosis of pancreatic tumor. Am J Gastroenterol 2002;97:2263-70. 23. Takahashi K, Yamao K, Okubo K, et al. Differential diagnosis of pancreatic cancer and focal pancreatitis by using EUS-guided FNA. Gastrointest Endosc 2005;61:76-9. 24. Ogura T, Yamao K, Sawaki A, et al. Clinical impact of K-ras mutation analysis in EUS-guided FNA specimens from pancreatic masses. Gastrointest Endosc 2012;75:769-74. 25. Visani M, De Biase D, Polifemo AM, et al. EUS-FNA for Kras mutational analysis in pancreatic lesions, preliminary results using three different techniques: Sanger sequencing, next generation sequencing and realtime PCR. Dig Liver Dis 2012;44:S65. 26. Jones S, Zhang X, Parsons DW, et al. Core signaling pathways in human pancreatic cancers revealed by global genomic analyses. Science 2008;321:1801-6. 27. Hruban RH, van Mansfeld AD, Offerhaus GJ, et al. K-ras oncogene activation in adenocarcinoma of the human pancreas: a study of 82 carcinomas using a combination of mutant-enriched polymerase chain
Figure 11. Algorithm for the management of solid pancreatic mass. During EUS-FNA, an additional pass should be performed for K-ras testing in case of inconclusive diagnosis. The result of K-ras gene analysis should then be considered in the clinical context and compared with all other diagnostic ?ndings. CA, carbohydrate antigen.
www.giejournal.org Volume 78, No. 4 : 2013 GASTROINTESTINAL ENDOSCOPY 607
Fuccio et al EUS-FNA and K-ras testing for pancreatic solid mass diagnosis
reaction analysis and allele-specific oligonucleotide hybridization. Am J Pathol 1993;143:545-54. 28. Hingorani SR, Tuveson DA. Ras redux: rethinking how and where Ras acts. Curr Opin Genet Dev 2003;13:6-13. 29. Tada M, Ohashi M, Shiratori Y, et al. Analysis of K-ras gene mutation in hyperplastic duct cells of the pancreas without pancreatic disease. Gastroenterology 1996;110:227-31. 30. Moskaluk CA, Hruban RH, Kern SE. p16 and K-ras gene mutations in the intraductal precursors of human pancreatic adenocarcinoma. Cancer Res 1997;57:2140-3. 31. Dumonceau JM, Polkowski M, Larghi A, et al. Indications, results, and clinical impact of endoscopic ultrasound (EUS)-guided sampling in gastroenterology: European Society of Gastrointestinal Endoscopy (ESGE) Clinical Guideline. Endoscopy 2011;43:897-912. 32. Nicaud M, Hou W, Collins D, et al. The utility of repeat endoscopic ultrasound-guided fine needle aspiration for suspected pancreatic cancer. Gastroenterol Res Pract 2010;2010:268-90. 33. Eloubeidi MA, Varadarajulu S, Desai S, et al. Value of repeat endoscopic ultrasound-guided fine needle aspiration for suspected pancreatic cancerJ Gastroenterol Hepatol 2008;23:567-70. 34. DeWitt J, McGreevy K, Sherman S, et al. Utility of a repeated EUS at a tertiary-referral center. Gastrointest Endosc 2008;67:610-9. 35. Poruk KE, Gay DZ, Brown K, et al. The clinical utility of CA 19-9 in pancreatic adenocarcinoma: diagnostic and prognostic updates. Curr Mol Med 2013;13:340-51. 36. Schoedel KE, Finkelstein SD, Ohori NP. K-Ras and microsatellite marker analysis of fine-needle aspirates from intraductal papillary mucinous neoplasms of the pancreas. Diagn Cytopathol 2006;34:605-8. 37. Sawhney MS, Devarajan S, O’Farrel P, et al. Comparison of carcinoembryonic antigen and molecular analysis in pancreatic cyst fluid. Gastrointest Endosc 2009;69:1106-10.
38. Khalid A, Zahid M, Finkelstein SD, et al. Pancreatic cyst fluid DNA analysis in evaluating pancreatic cysts: a report of the PANDA study. Gastrointest Endosc 2009;69:1095-102. 39. Rockacy MJ, Zahid M, McGrath KM, et al. Association between KRAS mutation, detected in pancreatic cyst fluid, and long-term outcomes of patients. Clin Gastroenterol Hepatol 2013;11:425-9. 40. Khalid A, McGrath KM, Zahid M, et al. The role of pancreatic cyst fluid molecular analysis in predicting cyst pathology. Clin Gastroenterol Hepatol 2005;3:967-73. 41. Garud SS, Willingham FF. Molecular analysis of cyst fluid aspiration in the diagnosis and risk assessment of cystic lesions of the pancreas. Clin Transl Sci 2012;5:102-7. 42. De Angelis C, Daniele L, Balbo-Mussetto A, et al. Feasibility and diagnostic accuracy of K-ras mutation analysis in fine needle aspirate under EUS guidance of pancreatic masses. Dig Liver Dis 2010;42S:S63. 43. Lee YJ, Choi Y-L, Lee KH, et al. Clinical utility if K-ras mutation analysis with washing fluid after EUS-FNA for pancreatic cancer diagnosis [abstract]. Gastrointest Endosc 2012;75:AB201. 44. Allegra CJ, Jessup JM, Somerfield MR, et al. American Society of Clinical Oncology provisional clinical opinion: testing for KRAS gene mutations in patients with metastatic colorectal carcinoma to predict response to anti-epidermal growth factor receptor monoclonal antibody therapy. J Clin Oncol 2009;27:2091-6. 45. Plesec TP, Hunt JL. KRAS mutation testing in colorectal cancer. Adv Anat Pathol 2009;16:196-203. 46. Longnecker DS, Terhune PG. What is the true rate of K-ras mutation in carcinoma of the pancreas? Pancreas 1998;17:323-4. 47. Wei S, Liang Z, Gao J, et al. Patterns of K-ras codon 12 and 13 mutations found in pancreatic adenocarcinoma of 30 Chinese patients by microdissection, PCR and direct sequencing. J Gastroenterol Hepatol 2005;20:67-72.
GIE on Twitter GIE now has a Twitter account. Followers will learn when the new issues are posted and will receive up-to-the-minute news as well as links to author interviews, podcasts, and articles. Search on Twitter for @GIE_Journal and follow all of GIE’s tweets.
608 GASTROINTESTINAL ENDOSCOPY Volume 78, No. 4 : 2013 www.giejournal.org
EUS-FNA and K-ras testing for pancreatic solid mass diagnosis Fuccio et al
PLACE THIS ORDER OR A SIMILAR ORDER WITH US TODAY AND GET AN AMAZING DISCOUNT