Research in Veterinary Science xxx (2012) xxx–xxx

Contents lists available at SciVerse ScienceDirect

Research in Veterinary Science

journal homepage: www.elsevier .com/locate / rvsc

A simple and rapid immunocytochemical technique for detectionof cytokeratin, vimentin, and S-100 protein in veterinary diagnostic cytology

Mariko Sawa a, Akira Yabuki a,⇑, Noriaki Miyoshi b, Kou Arai c, Osamu Yamato a

a Laboratory of Clinical Pathology, Department of Veterinary Medicine, Kagoshima University, Kagoshima, Japanb Laboratory of Veterinary Pathology, Department of Veterinary Medicine, Kagoshima University, Kagoshima, Japanc Kagoshima University Veterinary Teaching Hospital, Kagoshima University, Kagoshima, Japan

a r t i c l e i n f o a b s t r a c t

Article history:Received 28 October 2011Accepted 25 March 2012Available online xxxx

Keywords:CytokeratinCytologyRapid immunocytochemistryS-100 proteinVimentin

0034-5288/$ - see front matter � 2012 Elsevier Ltd. Ahttp://dx.doi.org/10.1016/j.rvsc.2012.03.013

⇑ Corresponding author. Address: Laboratory of ClinVeterinary Medicine, Kagoshima University, 1-21-240065, Japan. Tel./fax: +81 99 285 3561.

E-mail address: yabu@agri.kagoshima-u.ac.jp (A. Y

Please cite this article in press as: Sawa, M., etprotein in veterinary diagnostic cytology. Res. V

The objective of this study was to establish a simple and rapid immunocytochemical technique that canbe used in veterinary diagnostic cytology. Air-dried impression smears were collected from caninetumors. Samples of epithelial tumors, mesenchymal tumors, and malignant peripheral nerve sheathtumors and melanomas were used for detection of cytokeratin, vimentin, and S-100 protein, respectively.The labeled streptavidin–biotin system was used in the present study. Optimal fixation was determinedusing standard immunocytochemical procedures, and acetone fixation was found to be the most effec-tive. Optimal concentrations of primary and secondary antibodies were determined at a preset 5-minincubation. Omission of H2O2 treatment, shortening the time for blocking and labeled-streptavidin incu-bation, and simplifying washing did not decrease immunopositive intensities or enhance false-positivereactions. The described rapid protocol requires approximately 45 min without the use of any specialequipment.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Immunocytochemistry (ICC) has been an indispensable tech-nique for diagnostic cytology in human medicine over the last20 years (Skoog and Tani, 2011). This advanced diagnostic tech-nique has recently been used in veterinary cytology (Höinghauset al., 2007, 2008), and it provides additional information for cyto-morphological criteria by determining the origin of anaplastictumor cells such as those originating from poorly differentiatedcarcinomas, sarcomas, and melanomas; correct identification ofthese cells can be impossible using routine cytomorphology. Forexample, detection of the intermediate filament proteins (cytoker-atin and vimentin) helps determine whether tumor cells have orig-inated from epithelial or mesenchymal cells (Ramos-Vara et al.,2010). Detection of S-100 protein may help to discriminate theneurogenic or melanocytic tumors (Ramos-Vara et al., 2010).

In human medicine, ICC can be used not only to classify tumorsof unknown origin but also to detect therapeutic markers or infec-tious agents (Skoog and Tani, 2011). In patients with breast cancer,intraoperative ICC examination of sentinel lymph nodes is per-formed using an ultra-rapid technique (Salem et al., 2002; Francz

ll rights reserved.

ical Pathology, Department ofKorimoto, Kagoshima 890-

abuki).

al. A simple and rapid immunoet. Sci. (2012), http://dx.doi.or

et al., 2011). In certain neoplasms, ICC has shown some promisein differentiating benign and malignant tumors; it was also usefulfor preoperative evaluation (Ersoz et al., 2008; Ligato et al., 2008;Raggio et al., 2010). ICC is considered a powerful tool that assistscytomorphological diagnosis.

In veterinary medicine, ICC is not fully utilized in diagnosticcytology, although most veterinary clinical pathologists know ofits advantage and are skilled in ICC. ICC is performed based onthe procedure of immunohistochemistry (IHC). However, theIHC-based standard procedure may be inconvenient for most busyclinical pathologists because it requires many steps and is timeconsuming. Therefore, establishment of a simple and rapid ICC pro-cedure is required for ICC to be widely utilized, not only in diagnos-tic cytology but also in clinical and clinicopathological research inveterinary medicine.

The aim of the present study was to establish a simple and rapidICC procedure for detection of cytokeratin, vimentin, and S-100protein in canine tumor samples by modifying the standard ICCtechnique.

2. Materials and methods

2.1. Sample collections

Impression smears were prepared from surgical biopsy speci-mens that were obtained from dogs admitted to Kagoshima

cytochemical technique for detection of cytokeratin, vimentin, and S-100g/10.1016/j.rvsc.2012.03.013

2 M. Sawa et al. / Research in Veterinary Science xxx (2012) xxx–xxx

University Veterinary Teaching Hospital, Japan. All biopsy speci-mens were histopathologically diagnosed by a veterinary patholo-gist. Imprinted slides were thoroughly air-dried and stored at�30 �C until use. Epithelial specimens, including mammary ana-plastic carcinomas (n = 2), thyroid adenocarcinoma (n = 1), adeno-squamous carcinoma (n = 1), perianal gland adenoma (n = 1), andtransitional cell carcinomas (n = 2) were collected and these spec-imens were used for detection of cytokeratin. Mesenchymal spec-imens were collected from chondrosarcoma (n = 1), osteosarcoma(n = 1), hemangiosarcoma (n = 1), fibrosarcoma (n = 1), and myxo-sarcoma (n = 1), and these specimens were used for detectingvimentin. Specimens from poorly pigmented melanoma (n = 1)and malignant peripheral nerve sheath tumor (n = 1) were also col-lected and used for detecting S-100 protein.

2.2. Reagents

Immunocytochemistry was performed using a peroxidase-la-beled streptavidin–biotin system (LSAB, DakoCytomation, Glost-rup, Germany). Mouse monoclonal anti-cytokeratin (clone AE1/AE3, DakoCytomation), mouse monoclonal anti-vimentin (cloneV9, Thermo Fisher Scientific, Fremont, CA, USA), and rabbit poly-clonal anti-S-100 protein (cat. No. Z0311, DakoCytomation) wereused as primary antibodies. Biotin-labeled goat anti-mouseimmunoglobulin (IgG, Vector Laboratories, Burlingame, CA, USA)and goat anti-rabbit IgG (Vector Laboratories) were used as sec-ondary antibodies. All antibodies were diluted in 1% bovine serumalbumin (BSA)/10 mM phosphate-buffered saline (PBS). A 3% BSA/PBS solution was used as blocking solution. Reactivity was deter-mined using a 3,30-diaminobenzidine (DAB) system (Merck,Darmstadt, Germany), and Mayer’s hematoxylin was used forcounterstaining. For negative controls, non-immunized mouseIgG (DakoCytomation) and non-immunized rabbit IgG (R&D Sys-tems, Minneapolis, MN, USA) were used instead of primaryantibodies.

2.3. Determination of the optimum fixative

Acetone (1 min, 4 �C), 10% neutral buffered formalin (NBF;15 min, room temperature), and 95% ethanol (15 min, room tem-perature) were used as trial fixatives in the present study. Thedetection of cytokeratin, vimentin, and S-100 protein was testedin these 3 fixatives using standard ICC procedures.

Standard ICC in our laboratories was as follows: (1) thoroughlyair-dry the sample with cold air from a drier, (2) wash in distilledwater, (3) treat with 3% H2O2 for 30 min, (4) wash in PBS, (5)block for 60 min, (6) incubate overnight at 4 �C with primary anti-bodies (1:200 dilution for cytokeratin and vimentin; 1:400 dilu-tion for S-100 protein), (7) wash in PBS, (8) incubate for 30 minwith secondary antibodies (1:200 dilution), (9) wash in PBS, (10)incubate for 15 min with LSAB (ready-to-use), (11) wash in PBS,(12) react with DAB-H2O2 for 5 min, (13) stop the reaction by rins-ing in cold distilled water, and (14) counterstain with Mayer’shematoxylin.

Table 1Immunopositive signals and non-specific background staining for detection of cytokeratin

Assessment Acetone 10% NBF

Cytokeratin Vimentin S-100 Cytokera

Immunoreactivity 3+ 3+ 2+ 3+Background staining 1+ 1+ 1+ 1+

NBF: neutral buffered formalin, 1+: weak, 2+: moderate, 3+ strong.

Please cite this article in press as: Sawa, M., et al. A simple and rapid immunoprotein in veterinary diagnostic cytology. Res. Vet. Sci. (2012), http://dx.doi.or

2.4. Modification for rapid ICC

After determining the optimal fixative, the incubation time wasshortened and staining procedures were simplified. The followingincubations with antibodies and LSAB were performed at 37 �C.The incubation with antibodies was preset at 5 min, and dilutionsof primary antibodies were fixed at 1:10, 1:20, 1:50, 1:100; and1:200 for cytokeratin and vimentin and at 1:100, 1:200, and1:400 for S-100 protein. Similarly, dilutions of secondary antibod-ies were fixed at 1:150, 1:450, and 1:750 with 5-min incubation. Ifthe immunopositive signal in the best dilution was weaker thanthat of the standard ICC, incubation time was additionally testedat 10 and 15 min. The LSAB was a ready-to-use solution, and itsincubation time was set at 5, 10, and 15 min. The best dilutionsof antibodies and the best incubation time of LSAB were deter-mined by comparing the intensity of immunopositive signals tothe standard procedure. Staining procedures were further short-ened and simplified as follows: H2O2 treatment was omitted, andthe blocking time was shortened to 5–10 min. PBS washing wasperformed using a direct stream from a washing bottle for approx-imately 10 s.

3. Results

3.1. Determination of the optimum fixative

The intensities of immunoreactivity and background stainingfrom 3 fixatives are summarized in Table 1. For detection of cyto-keratin, clear and strong immunopositive signals were observed inall fixatives (Fig. 1A–C). For detection of vimentin, strong immuno-positive signals were observed in acetone fixation (Fig. 1D–F). Fordetection of S-100 protein, moderate immunopositive signals wereobserved in all fixatives (Fig. 1G–I). Background staining was rela-tively lower in acetone than in other fixatives.

3.2. Dilution and incubation time of antibodies

The investigations described below were performed in acetonefixation. With the anti-cytokeratin and anti-vimentin antibodies,dilutions of 1:10, 1:20, 1:50, 1:100, and 1:200 were tested at 5-min incubation, and clear and intense immunopositive signalswere observed at 1:50 dilution in cytokeratin and at 1:100 invimentin. For anti-S-100 protein antibody, dilutions of 1:100,1:200, and 1:400 were tested at 5 min. Subsequently, 10- and15-min incubations were tested, and clear immunopositive signalswere observed at 1:100 dilution for 10-min incubation.

For the secondary antibody, dilutions of 1:150, 1:450, and 1:750were tested at 5-min incubation, and clear immunopositive signalswere observed at 1:150 dilution. For LSAB, 5 min of incubation wassufficient to detect the signals. Thus, optimal conditions of re-agents were determined as follows: anti-cytokeratin, 1:50 dilutionand incubated for 5 min; anti-vimentin, 1:100 dilution and incu-bated for 5 min; anti-S-100 protein, 1:100 dilution and incubatedfor 10 min; secondary antibody, 1:150 dilution and incubated for5 min; and LSAB, ready-to-use and incubated for 5 min.

, vimentin, and S-100 protein in the 3 different fixatives.

95% Ethanol

tin Vimentin S-100 Cytokeratin Vimentin S-100

1+ 2+ 3+ 1+ 2+2+ 2+ 1+ 2+ 2+

cytochemical technique for detection of cytokeratin, vimentin, and S-100g/10.1016/j.rvsc.2012.03.013

Table 2Procedure of rapid immunocytology for cytokeratin, vimentin, and S-100 protein.

1 Fixation in acetone for 1 minute at 4 �C2 Washing with PBSa

3 Blocking with 3% BSA/PBS for 5–10 min4 Incubation with the primary antibodies for 5-10 min at 37 �Cb

5 Washing with PBSa

6 Incubation with the biotin-labeled second antibodies for 5 min at 37 �C7 Washing with PBSa

8 Incubation with peroxidase-labeled streptavidin for 5 min at 37 �C9 Washing with PBSa

10 Incubate with DAB for 5 min11 Termination of the reaction in cold distilled water12 Counterstaining with Mayer’s hematoxyline

a 10 s with direct stream from a bottle.b 5 min for anti-cytokeratin and anti-vimentin antibodies, and 10 min for anti-S-

100 protein antibody.

Fig. 1. Standard immunocytochemistry for detection of cytokeratin, vimentin, and S-100 protein. (A–C) cytokeratin, mammary anaplastic carcinoma. (D–F) vimentin,osteosarcoma. (G–I) S-100 protein, malignant peripheral nerve sheath tumor. (A, D, and G) aceton fixation. (B, E, and H) 10% neutral buffered formalin (NBF) fixation. (C, F, andI) 95% ethanol fixation. Bars = 30 lm. Counterstain, Mayer’s hematoxylin.

M. Sawa et al. / Research in Veterinary Science xxx (2012) xxx–xxx 3

3.3. Further shortening and simplification of protocol

Omission of the H2O2 treatment, shortening of blocking time,and a brief PBS washing did not enhance false-positive reactions.

3.4. Rapid ICC for cytokeratin, vimentin, and S-100 protein

The rapid ICC procedure established in the present study isshown in Table 2. Intensities and specificities of immunosignalsfor cytokeratin, vimentin, and S-100 protein obtained using this ra-pid protocol were sufficient compared to those of standard proto-cols (Fig. 2). In addition, damage to cellular morphology and thedegree of exfoliation of cells were apparently mild in rapid ICCcompared to those in standard ICC in most of the samples (Fig. 2).

Please cite this article in press as: Sawa, M., et al. A simple and rapid immunoprotein in veterinary diagnostic cytology. Res. Vet. Sci. (2012), http://dx.doi.or

4. Discussion

The superiority of acetone fixation in ICC has been demon-strated in human lymph node cytology (Salem et al., 2003). In vet-erinary cytology, a recent report successfully demonstrated clearimmunopositive signals for vimentin, S-100 protein, and melan-Ain canine mesenchymal tumors using acetone fixation (Höinghauset al., 2008). In the present study, the suitability of acetone fixationwas demonstrated by a comparative analysis using 3 different fix-atives. Since exfoliation of cells from the slides and fragmentationof cytoplasm during the staining process are disadvantages of ace-tone fixation in ICC (Valli et al., 2009), short-time fixation was usedin the present study. One-minute fixation in acetone was sufficientfor the detection of cytokeratin, vimentin, and S-100 protein fromair-dried cytological samples.

In the present study, it was demonstrated that NBF and 95% eth-anol can be used for the detection of cytokeratin and S-100 proteinbut not vimentin. NBF is the most popular fixative in IHC, and itsusefulness for ICC was pointed out in veterinary cytology (Valliet al., 2009). Ninety-five percent ethanol is a commonly used fixa-tive for ICC in human cytology (Shidham et al., 2000; Maeda et al.,2005). However, the present study found no advantage betweenNBF and 95% ethanol compared to acetone as a fixative for thedetection of cytokeratin, vimentin, and S-100 protein.

Shortening of incubation time is a necessary step for accom-plishing rapid ICC. Several methods for shortening the incubationtime of antibodies have been previously reported, and a very sim-ple method has been used in a rapid intraoperative ICC in humanmedicine (Maeda et al., 2005; Francz et al., 2011). Briefly, thismethod requires only highly concentrated antibody solutions,and a few minutes of incubation is sufficient for detecting antigenssuch as cytokeratin and cluster of differentiation (CD) antigens.Although the present modification is based on this highly concen-

cytochemical technique for detection of cytokeratin, vimentin, and S-100g/10.1016/j.rvsc.2012.03.013

Fig. 2. Comparison of rapid immunocytochemistry and standard immunocytochemistry. (A and B) cytokeratin, mammary anaplastic carcinoma. (C and D) cytokeratin,thyroid adenocarcinoma. (E and F) vimentin, myxosarcoma. (G and H) vimentin, fibrosarcoma. (I and J) S-100 protein, poorly pigmented melanoma. (A, C, E, G, and I) rapidimmunocytochemistry. (B, D, F, H, and J) standard immunocytochemistry. Bars = 30 lm. Counterstain, Mayer’s hematoxylin.

4 M. Sawa et al. / Research in Veterinary Science xxx (2012) xxx–xxx

trated antibody technique, non-economic consumption of anti-body solution and the possibility of non-specific background stain-ing were regarded as disadvantages of this simple method. Sincethe present study emphasized easy handling, low backgroundstaining, and observation of clear, specific immunopositive signals,the qualities of cytokeratin, vimentin, and S-100 protein immu-noreactivities were evaluated at different concentrations of anti-bodies for 5 min at 37 �C. This incubation time and temperaturewas determined on the basis of a previous report regarding the

Please cite this article in press as: Sawa, M., et al. A simple and rapid immunoprotein in veterinary diagnostic cytology. Res. Vet. Sci. (2012), http://dx.doi.or

ICC of human lymph nodes (Salem et al., 2003). Briefly, this reportevaluated the immunoreactivity against cytokeratin using en-hanced polymer one step (EPOS) method and demonstrated thatincubation with antibody for 5 min at 37 �C resulted in acceptableto strong immunoreactivity. In the present study, cytokeratin wasclearly detected in 1:50 (3.6 lg/ml) dilution of antibody, whichcorresponded to the manufacturer’s recommended concentration.Vimentin was clearly detected in 1:100 (2.0 lg/ml) dilution, whichcorresponded to the lowest concentration recommended by the

cytochemical technique for detection of cytokeratin, vimentin, and S-100g/10.1016/j.rvsc.2012.03.013

M. Sawa et al. / Research in Veterinary Science xxx (2012) xxx–xxx 5

manufacturer (1:50–1:100; 2.0–4.0 lg/ml). The optimal concen-tration of secondary antibodies was determined to be 1:150(10.0 lg/ml) dilution, which corresponded to the highest concen-tration recommended by the manufacturer (1:150–1:750; 2.0–10.0 lg/ml). These findings demonstrated that 5-min incubationoften accomplished rapid and sufficient antigen–antibody reac-tions without excessively concentrated antibodies. On the otherhand, highly concentrated primary antibody was required to detectacceptable signals for S-100 protein, and the 1:100 dilution(10.25 lg/ml) used in the present study corresponds to 4 timesthe dilution recommended by the manufacturer (1:400; 2.56 lg/ml). In addition, the incubation time required to detect the accept-able immunoreactivity was 10 min. The anti-S-100 protein anti-body used in the present study was a rabbit polyclonal antibody,in contrast to the monoclonal antibodies against cytokeratin andvimentin; therefore, it was suspected that a monoclonal antibodymight be most suitable for the present rapid ICC method.

LSAB was chosen as a detection system in the present immun-ocytology because of its reliability, wide application, and low cost.Although ready-to-use solution was used in the present study, 5-min incubation was enough to perform a sufficient biotin-strepta-vidin interaction.

H2O2 treatment for the removal of endogenous peroxidaseactivity is often omitted during intraoperative rapid ICC in humanmedicine (Katayama et al., 2004). In the present study, the effect ofomission of H2O2 treatment was evaluated, and no enhancement offalse-positive reactions was found after the omission. However, ifthe samples contain many endogenous peroxidase-rich cells likemacrophages, H2O2 treatment might be required for accurate inter-pretation of the immunosignals. In such cases, a mild reagent (forexample 0.03% H2O2) would be preferable because cellular mor-phology in smear cytology may be damaged by 3% H2O2 treatment(Key, 2006). Actually, in the present study, damage to the cellularmorphology and exfoliation of cells were observed in the standardICC using 3% H2O2 treatment.

In conclusion, the present study established a rapid ICC methodfor air-dried samples by modifying the standard protocol. This ra-pid method enabled detection of cytokeratin, vimentin, and S-100protein in canine tumor cells, and satisfactory immunopositive sig-nals could be observed within 45 min. In this technique, simplicityand easy handling are emphasized because no special equipment isrequired to perform this rapid ICC. In small animal medicine, sev-eral antibodies are currently available for ICC (Ramos-Vara et al.,2010), and the rapid ICC method established in the present studymay be used for many other antibodies. However, the type offixative and the dilution and incubation time of the antibody

Please cite this article in press as: Sawa, M., et al. A simple and rapid immunoprotein in veterinary diagnostic cytology. Res. Vet. Sci. (2012), http://dx.doi.or

should be determined in advance. In addition, the tissue of originin the sample and type and degree of inflammation should bepre-checked to determine whether endogenous peroxidase block-ing is necessary.

References

Ersoz, S., Sert, H., Yandi, M., Erem, C., Mungan, S., Ersoz, H.O., Cobanoglu, U.,Hacihasanoglu, A., 2008. The significance of galectin-3 expression in theimmunocytochemical evaluation of thyroid fine needle aspiration cytology.Pathology and oncology Research 14, 457–460.

Francz, M., Egervari, K., Szollosi, Z., 2011. Intraoperative evaluation of sentinellymph nodes in breast cancer: comparison of frozen sections, imprint cytologyand immunocytochemistry. Cytopathology 22, 36–42.

Höinghaus, R., Hewicker-Trautwein, M., Mischke, R., 2007. Immunocytochemicaldifferentiation of neoplastic and hyperplastic canine epithelial lesions incytologic imprint preparations. The Veterinary Journal 173, 79–90.

Höinghaus, R., Hewicker-Trautwein, M., Mischke, R., 2008. Immunocytochemicaldifferentiation of canine mesenchymal tumors in cytologic imprintpreparations. Veterinary Clinical Pathology 37, 104–111.

Katayama, H., Maeda, S., Hosone, M., Hara, H., Sanno, N., Shimura, T., Yokoyama, M.,Naito, Z., 2004. A case of glioblastoma (gemistocytic type) immunostainingefficacy in intraoperative diagnosis. Journal of the Japanese Society of ClinicalCytology 43, 331–334 (in Japanese with English abstract).

Key, M., 2006. Methods of immunocytology for slide-based cellular analysis. In: Key,M. (Ed.), Immunocytochemical Staining Methods, 4th ed. Dako, California, pp.95–101.

Ligato, S., Mandich, D., Cartun, R.W., 2008. Utility of glypican-3 in differentiatinghepatocellular carcinoma from other primary and metastatic lesions in FNAof the liver: an immunocytochemical study. Modern Pathology 21,626–631.

Maeda, S., Yokoyama, M., Naito, Z., 2005. Rapid cytological diagnosis correlated withclinical medicine. NIHON IKA DAIGAKU IGAKUKAI ZASSH 1, 102–109 (inJapanese with English abstract).

Raggio, E., Camandona, M., Solerio, D., Martino, P., Franchello, A., Orlandi, F.,Gasparri, G., 2010. The diagnostic accuracy of the immunocytochemicalmarkers in the pre-operative evaluation of follicular thyroid lesions. Journalof Endocrinological Investigation 33, 378–381.

Ramos-Vara, J.A., Avery, A.C., Avery, P.R., 2010. Advanced Diagnostic Techniques. In:Raskin, R.E., Meyer, DE., (Eds.), Canine and Feline Cytology: A Color Atlas andInterpretation Guide, 2nd ed. Saunders, Missouri, pp. 395–437.

Salem, A.A., Douglas-Jones, A.G., Sweetland, H.M., Newcombe, R.G., Mansel, R.E.,2002. Evaluation of axillary lymph nodes using touch imprint cytology andimmunohistochemistry. The British Journal of Surgery 89, 1386–1389.

Salem, A.A., Douglas-Jones, A.G., Sweetland, H.M., Mansel, R.E., 2003. Intraoperativeevaluation of axillary sentinel lymph nodes using touch imprint cytology andimmunohistochemistry: I. Protocol of rapid immunostaining of touch imprints.European Journal of Surgical Oncology 29, 25–28.

Shidham, V.B., Lindholm, P.F., Kajdacsy-Balla, A., Chang, C.C., Komorowski, R., 2000.Methods of cytologic smear preparation and fixation. Effect on theimmunoreactivity of commonly used anticytokeratin antibody AE1/AE3. ActaCytologica 44, 1015–1022.

Skoog, L., Tani, E., 2011. Immunocytochemistry: an indispensable technique inroutine cytology. Cytopathology 22, 215–229.

Valli, V., Peters, E., Williams, C., Shipp, L., Barger, A., Chladny, J., Hoffmann, W., 2009.Optimizing methods in immunocytochemistry: one laboratory’s experience.Veterinary Clinical Pathology 38, 261–269.

cytochemical technique for detection of cytokeratin, vimentin, and S-100g/10.1016/j.rvsc.2012.03.013