Background Fine‐needle aspirate (FNA) cytology is often the first‐choice method for diagnosing gastrointestinal nodular lesions. The FNA material can be converted to histopathology specimens by a needle rinse cell block (NRCB) technique, allowing ancillary studies to refine the cytologic diagnosis. Despite use in human pathology, NRCB has never been applied to canine or feline gastrointestinal neoplasia. Objective This study described NRCB methodology and its diagnostic utility in specific cases of neoplastic gastrointestinal lesions. Methods Needle rinses with saline were performed after ultrasound‐guided FNAs of two intestinal lymphomas (canine and feline) and a canine gastrointestinal stromal tumor (GIST). The NRCB was prepared using the cell tube block technique and processed for paraffin embedding. Routine immunohistochemistry protocols (using CD3, PAX‐5, and Ki‐67 for lymphoma cases and vimentin, desmin, S‐100, and KIT markers for GIST) were applied to NRCB sections, and the results were compared with matched tissue biopsies. Results NRCBs with adequate cell numbers, preservation, and good separation of blood were obtained. The diagnosis and immunophenotyping were confirmed in both cases of lymphoma in NRCBs. In the GIST, the immunolabeling of the neoplastic cells in NRCB was completely concordant with the tissue biopsy. Conclusions The described methodology is suitable for veterinary settings, having few technical requirements and low invasiveness. The presented cases of gastrointestinal neoplasia highlight the utility of NRCBs as a platform to conduct ancillary studies and refine the cytologic diagnosis.

Immunolabeling on Romanowsky-stained cytology (RSC) slides can be used, although there is limited evidence of its suitability for phenotyping canine and feline lymphomas. A comparison with matched cell blocks (CB) is missing. Immunolabeling on RSC and CB was compared for lymphoid markers (CD3 and PAX5) in 53 lymphomas and 4 chylous effusions from dogs and cats. The influence of pre-analytical variables (species, time of archive, type of specimens and coverslipping) and the interobserver agreement among the 2 observers was assessed. Fewer CD3+ lymphocytes were identified in RSC, while the PAX5 positivity by RSC and CB had a substantial agreement. Immunodetection of CD3 and the diagnosis of a T-cell population on RSC was more difficult. Lower intensity and higher background were noted in RSC. Immunophenotyping was inconclusive in 54% RSC and 19% CB. The interobserver reproducibility of immunophenotyping on CB was substantial, being higher than in RSC. The immunolabeling performance on the RSC of effusion and feline samples was unsatisfactory. The detection of lymphoid markers, especially membranous antigens in retrospective RSC, is affected by the pre-analytical variables: species, time of the archive, and type of specimens. CB are a more consistent type of sample for immunophenotyping purposes.