Article Text

Download PDFPDF

EGFR mutations detection on liquid-based cytology: is microscopy still necessary?
  1. Umberto Malapelle1,
  2. Nicla de Rosa2,
  3. Claudio Bellevicine1,
  4. Danilo Rocco2,
  5. Fabiana Vitiello2,
  6. Franco Vito Piantedosi2,
  7. Alfonso Illiano2,
  8. Oscar Nappi3,
  9. Giancarlo Troncone1,4
  1. 1Dipartimento di Scienze Biomorfologiche e Funzionali, Università di Napoli Federico II, Naples, Italy
  2. 2AORN Vincenzo Monaldi, Naples, Italy
  3. 3AO Antonio Cardarelli, Naples, Italy
  4. 4CEINGE–Biotecnologie Avanzate, Naples, Italy
  1. Correspondence to Professor Giancarlo Troncone, Dipartimento di Scienze Biomorfologiche e Funzionali, Università di Napoli Federico II, via Sergio Pansini 5, Napoli I-80131, Italy; giancarlo.troncone{at}


Currently, there is a trend towards an increasing use of liquid-based cytology (LBC) to diagnose non–small cell lung cancer. In this study, to detect epidermal growth factor receptor mutations, different molecular techniques were applied to LBC samples with and without laser capture microdissection (LCM). In 58 LBCs, DNA was extracted twice. One sample was obtained directly from CytoLyt solution, whereas the other DNA sample was derived after slide preparation and LCM of Papanicolaou-stained cells. The rate of mutant cases obtained by direct sequencing was discordant between CytoLyt-derived (10.3%) and LCM-derived (17.2%) DNA. However, the same mutant rate (17.2%) was achieved on the matched samples by high-resolution melting analysis, fragment and TaqMan assays. Thus, LCM and direct sequencing may be replaced by more sensitive non-sequencing methods directly performed on CytoLyt-derived DNA, an easier and faster approach to improve epidermal growth factor receptor testing standardisation on LBCs.

  • Cancer research
  • cytology
  • molecular pathology
  • lung
  • molecular oncology
  • thyroid cancer
  • morbid anatomy
  • lung cancer
  • colorectal cancer
  • thyroid

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Testing for epidermal growth factor receptor (EGFR) mutations is current practice in non–small cell lung cancer (NSCLC) therapeutic management.1 In the advanced stages of NSCLC, histological samples are often not available1; thus, cytological samples are being increasingly used to genotype NSCLC for EGFR mutations.1 However, several technical issues make difficult to standardise EGFR testing on cytology.2 Fixation may greatly differ among institutions, and cytological samples may be prepared as direct smears, cytospins or cell blocks2; consequently, EGFR testing protocols optimised on a given preparation type are not consistently reproducible on a different cytological biospecimen.2 The variability degree is further increased by the different modalities of cancer cell enrichment and by the analytical sensitivity of the method adopted to detect EGFR mutation.1

Liquid-based cytology (LBC) standardises sample collection and processing.3 Instead of being smeared, cells are rinsed into a liquid preservative collection medium (CytoLyt) and processed on automated devices to prepare a monolayered LBC slide. In a recent study of ours, EGFR mutations were reliably detected, even when only a few cells were present on an LBC slide.3 To achieve this sensitivity, direct sequencing was performed after laser capture microdissection (LCM) of cancer cells. Conversely, direct sequencing gave false-negative results when DNA was directly extracted from CytoLyt.3 Since LCM requires special instruments and it is time consuming and costly, this study aimed to replace LCM and direct sequencing by more sensitive non-sequencing methods directly applied on CytoLyt-derived DNA.


Study samples and design

The molecular cytopathology laboratory at the University Federico II of Naples is a southern Italian large reference centre for treatment-predictive EGFR and KRAS mutation analysis.3 Between July 2010 and December 2011 a total of 560 NSCLC samples were referred to our laboratory from 12 different hospitals for EGFR testing. A large portion of cases (367/527) was represented by cytological specimens; these included LBCs (n=208), archival slides (n=127) and cell blocks (n=32). We focused on LBCs. Study protocol was approved (protocol 185/10) by the University of Napoli Federico II ‘Carlo Romano’ Ethics Committee. Special care was taken to select only those LBCs whose phial contained not <10 ml of residual CytoLyt (Cytyc UK, Crawley, UK) preservative solution. Fifty-eight samples were available; these included 22 cases, whose EGFR status had already been assessed by direct sequencing on both cell pellet- and LCM-derived DNA, as described in a previous study.3 Most (44/58) cases had been diagnosed as adenocarcinoma (ADC), one was a squamous cell carcinomas (SQCC) and 13 were NSCLC not otherwise specified. In 17 cases, NSCLC subtyping had also been based on thyroid transcription factor 1, p63, cytokeratin 5/6 and Napsin A immunostainings.

Study design is shown in figure 1; in each single case, DNA was extracted twice. One sample was directly obtained from CytoLyt solution, whereas the other DNA sample was derived after LCM of Papanicolaou ThinPrep slide, as previously described.3

Figure 1

Study design. Epidermal growth factor receptor (EGFR) mutation detection was carried out by direct sequencing, high-resolution melting analysis (HRMA), fragment and TaqMan assays both on DNA directly extracted from the CytoLyt solution and on that obtained from the laser capture microdissected cells. LBC, liquid-based cytology.

EGFR mutation testings

On both samples, exons 19 and 21 EGFR mutational status was assessed by direct sequencing and by high-resolution melting analysis (HRMA), as previously described.4 5 In addition, fragment assay for exon 19 and TaqMan assay for exon 21 were performed, as described by Rosell et al.6 To ensure data consistency, all assays were carried out in duplicate.


Exons 19 and 21 EGFR mutation detection on cell pellet-derived DNA

Direct sequencing identified EGFR mutations in 6/58 (10.3%) of the cases. These mutations occurred in four instances on exon 19 (E746_A750 deletion n=3; L747_A750 deletion n=1) and in two instances on exon 21 (L858R mutation n=2). All these mutations were also detected by the other methods, which in addition found four more mutant cases. As shown in table 1, the E746_A750 exon 19 deletion was detected by HRMA and by fragments assay in case 44 (figure 2). Similarly, three more cases (n. 17; n. 38; n. 42) harbouring the exon 21 L858R mutation were detected by the HRMA and TaqMan assays (figure 3).

Table 1

Description of the four discrepant test results

Figure 2

Case 44. (A) Cell pellet; E746_A750 exon 19 deletion was detected by fragments assay (arrow) by high-resolution melting analysis (difference plot) but not by direct sequencing. (B) Laser capture microdissection cells. The E746_A750 exon 19 deletion was detected by all the assays.

Figure 3

Case 17. (A) Cell pellet. The L858R exon 21 mutation was detected by TaqMan assays (allelic discrimination plot) and by high-resolution melting analysis (difference plot) but not by direct sequencing. (B) Laser capture microdissection cells. The L858R exon 21 mutation was detected by all the assays.

Exons 19 and 21 EGFR alteration detection on LCM-derived DNA

Direct sequencing on LCM cells yielded results identical to those obtained by the other methods. In particular, all were concordant in detecting in five instances an alteration on exon 19 (E746_A750 deletion n=4; L747_A750 deletion n=1). As far as exon 21 is concerned, the L858R mutation was identified on five LCM-derived DNAs by direct sequencing, HRMA and TaqMan assays (table 1).


This study validated an easier and faster approach to improve EGFR testing on LBC. LCM and direct sequencing may be replaced by more sensitive non-sequencing methods directly performed on CytoLyt-derived DNA. In current practice, the primary pathologist sends either an archival slide or a cell block to a reference laboratory for external testing. Several steps are necessary before the EGFR mutational assay. Our method omits the microscopic evaluation of cellular adequacy, the sample preparation for DNA extraction, involving either cell block sectioning or coverslip removal, and the manual or the laser-assisted microdissection of neoplastic cells. These steps are costly and time consuming and require the expertise of referral cytopathologist, who is not always included in the staff of the external molecular laboratory.

To reliably detect EGFR mutations on cytological samples, two approaches can be pursued; either LCM can be performed prior to direct sequencing3 4 or more sensitive mutation detection techniques can be exploited, without tumour cells enrichment.7–9 Here, we showed that both approaches can be applied to LBCs. The rate of EGFR mutant cases was 17.2%; this suggests that our method is highly sensitive, while the specificity was ensured by performing all assays in duplicate. Further prospective studies are warranted to assess whether an increased analytical sensitivity corresponds to an enhanced clinical sensitivity in order to better predict the response to EGFR antagonists treatment in NSCLC patients.

In conclusion, we propose a simple approach to standardise EGFR mutational testing on cytology, combining sensitive molecular techniques to the uniform method of sample collection and preparation ensured by LBC.

Take-home messages

  • Direct sequencing requires laser capture microdissection of cancer cells from Papanicolaou stained smears.

  • Instead, high-resolution melting analysis, fragment and TaqMan assays enable testing Cytolyt rather than smears.

  • Thus the microscopic evaluation of cells to be tested may be not longer needed.



  • Funding This work was partially supported by Astrazeneca (grant number ISSIRES0025).

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval Ethics approval was provided by Università degli Studi di Napoli Federico II, Comitato Etico per le attività biomediche “Carlo Romano” n. study 185/10.

  • Provenance and peer review Not commissioned; externally peer reviewed.