neoplasm that is thought to arise from the reserve cells of the transformation zone. Morphologically, SMILE is composed of immature stratified cells that display intracytoplasmic mucin or cytoplasmic vacuoles, but lack overt gland formation that is typically seen in AIS. SMILEs are commonly associated with HSIL, AIS and invasive carcinoma. Since the first description of histologic features of SMILE in 2000, there has been a lack of description of cytologic features of SMILE in the literature. We have retrospectively reviewed cytologic features of four cases of SMILE and have discussed its pitfalls in cervical cytology. Materials and Methods: Four cytologic cases from two patients with SMILE were identified in the hospital archives. The corresponding cytology (ThinPrep) and histology slides were reviewed and compared. Immunohistochemistry for p16 was used for evaluation on the histologic material. High-risk HPV (hr-HPV) status results, analyzed with Hybrid Capture 2, were documented. Results: Histologically, SMILEs showed less impressive nuclear changes with enlarged nuclei and increased nuclear/cytoplasmic ratio. The dysplastic epithelium retained mucin-producing columnar cells with prominent intracytoplasmic mucin. Mitotic figures were encountered, but apoptotic bodies were not seen. p16 immunostaining revealed both nuclear and cytoplasmic full thickness staining of the dysplastic epithelium. HrHPV testing was positive in all four cytology cases. Cytologically, crowded groups of atypical endocervical glandular cells were consistently seen in all four cases. These groups of atypical glandular cells showed slightly enlarged nuclei, nuclear overlapping, hyperchromasia and prominent nucleoli. The borders of these cell groups were relatively smooth. In three cases, LSIL-like atypical squamous cells were encountered. Original cytologic diagnoses were rendered as negative, ASC-US (x2) and LSIL. The glandular cell abnormality was not mentioned in any of the original cytologic diagnoses. Conclusions: Similar to the histology, cytologic features of SMILE are less dramatic and are often subtle. It lacks the typical cytologic changes seen in either AIS or HSIL. The consistent finding is that of relatively crowded groups of atypical glandular cells. The architectural abnormality is more prominent than the nuclear changes. The features are not typical for AIS and these atypical glandular cells could easily be interpreted as reactive given their overlapping features with those seen in repair. In addition, concurrent atypical squamous cells serve as a distracting factor in overlooking these atypical glandular cells. Our study emphasizes that careful review of crowded groups of glandular cells in hr-HPV positive women is absolutely critical. Based on our knowledge, this is the first description of the cytologic features of these lesions.