Analysis of microRNA Expression in Liquid-Based Cytology Samples may be Useful for Primary Lung Cancer Diagnosis

Yusuke Araki, MD; Koji Arihiro, MD, PhD; Kakuhiro Yamaguchi, MD, PhD; Shinjiro Sakamoto, MD, PhD; Yasushi Horimasu, MD, PhD; Takeshi Masuda, MD, PhD; Shintaro Miyamoto, MD, PhD; Taku Nakashima, MD, PhD; Hiroshi Iwamoto, MD, PhD; Kazunori Fujitaka, MD, PhD; Hironobu Hamada, MD, PhD; Noboru Hattori, MD, PhD


Am J Clin Pathol. 2021;156(4):644-652. 

In This Article

Abstract and Introduction


Objectives: Bronchoscopy is frequently performed for patients suspected of having lung cancer; however, we sometimes fail to make a definitive diagnosis, resulting in additional invasive testing. Many studies indicate that microRNAs (miRs) are abnormally expressed in cancers. We examined the diagnostic value of 4 miRs (miR-21, miR-31, miR-182, and miR-183) extracted from liquid-based cytology (LBC) samples and validated whether they were diagnostically useful.

Methods: We collected 18 surgically resected tissue samples and 136 LBC specimens obtained during bronchoscopic examination at Hiroshima University Hospital. We extracted RNA from these samples and compared the expression of 4 miRs by reverse transcription–quantitative polymerase chain reaction.

Results: We confirmed that expression of the 4 miRs was significantly higher in cancer tissues than in tumor-adjacent normal tissues. We examined the expression of these miRs in 125 (cancer cases, 83; noncancer cases, 42) of 136 cytologic samples. Expression of all 4 miRs was significantly higher in patients with lung cancer than in those without lung cancer. Among samples judged as benign or indeterminate, levels of these miRs were also significantly higher in patients with lung cancer than in those without lung cancer.

Conclusions: The analysis of miR expression in LBC samples might be helpful for primary lung cancer diagnosis.


Lung cancer is the leading cause of cancer deaths in the world. Bronchoscopy is useful for the diagnosis of lung cancer, so we perform bronchoscopy when we detect nodules, which are suspected as lung cancer, by chest imaging. However, we sometimes fail to collect adequate cytologic and/or biopsy samples because of tumor location or size. Consequently, it is hard for us to diagnose certain patients as having lung cancer despite the presence of a lung tumor identified by radiologic examination. Conversely, we usually use classification schemes in evaluating cytologic specimens. When it is difficult to differentiate atypical cells from carcinoma cells, we judge the cytologic specimen to be indeterminate. Therefore, some cases undergo surgical excision for histologic diagnosis or additional invasive examinations that can result in the diagnosis of benign lesions.

MicroRNAs (miRs) are small noncoding RNAs consisting of approximately 22 nucleotides that play important roles in the regulation of gene expression.[1] Many previous studies have revealed significant differences in miR expression between lung cancer tissue and noncancerous lung tissue[2,3] and in the diagnostic value of serum miRs.[4,5] However, few studies have been published regarding the diagnostic value of cytologic specimens. Xie et al[6] showed that expression of miR-21 in patients with cancer was clearly higher than that in patients who were cancer-free. Huang et al[7] showed that expression levels of miR-29a and miR-375 of bronchial brush specimens accurately differentiate small cell lung cancer from non–small cell lung cancer and that expression of miR-34a and miR-205 of such brush specimens accurately differentiates squamous cell carcinoma from adenocarcinoma. However, there have been no studies concerning the diagnosis of lung cancer or the validity or utility of profiles of miRs extracted from liquid-based cytology (LBC) samples collected by bronchoscopy.

We investigated whether the expression of miRs in cytologic samples could be useful as diagnostic biomarkers in patients with lung cancer. If we could demonstrate that miR expression levels in such specimens were useful for diagnosis, this approach might be applicable to routine samples. Previous studies have reported that miR-21 is upregulated in various human malignancies[8,9] and can modulate tumorigenesis, cell proliferation, and metastasis in various cancers. For example, miR-21 targets PTPN14 in intrahepatic cholangiocarcinoma[10] and 15-PGDH in gastric cancer.[11] As for lung cancer, miR-21 has been used as a diagnostic marker in many studies.[12] It has been reported that levels of miR-21 increase with oncogenic K-Ras activation and can modulate non–small cell lung cancer tumorigenesis[13] by targeting PTEN[14] and EZH2.[15] Furthermore, high levels of circulating miR-21 in serum have been revealed to be useful for diagnosis of lung cancer. Consequently, miR-21 is considered a promising biomarker in lung cancer.[16,17]

One of the various oncogenes involved in colon and uterine cervical cancer is miR-31.[18,19] For lung cancer, Liu et al[20] showed that miR-31 can regulate LATS2 and PPP2R2A in human lung cancer cell lines. Edmonds et al[21] showed that miR-31 is overexpressed in human lung adenocarcinoma and that overexpression of miR-31 independently correlates with poor survival and accelerates lung tumorigenesis in mice. Both miR-182 and miR-183 are members of the miR-183 cluster belonging to the polycistronic miRNA cluster located in the 5-kb region of human chromosome 7q32.3.[22] Both miR-182 and miR-183 have been reported to be upregulated in human malignancies.[23,24] Previous studies have shown that tissue and serum levels of miR-182 and miR-183 are elevated in patients with lung cancer compared with healthy controls.[25,26] It is known that miR-182 promotes lung cancer tumorigenesis by modulating pyruvate dehydrogenase kinase 4[27] and exerts an oncogenic role by targeting FBXW7 and FBXW11.[28] In addition, miR-183 has been reported to be one of various oncogenes by targeting EGR1 and promoting tumor cell migration.[29]

In this study, we examined 4 miRs—miR-21, miR-31, miR-182, and miR-183—in surgically resected cancerous and noncancerous lung tissues and in cytologic samples of lung cancers and noncancerous lesions, and we verified whether the accuracy of discrimination between benign and malignant cases can be improved by using the expression profiles of these 4 miRs extracted from cytologic samples.