To explore the application value of the Yokohama reporting system in rapid on-site evaluation (ROSE) of touch imprints from core-needle biopsy (CNB) of breast masses, and to evaluate the clinical significance of patient triage management based on ROSE classification according to this reporting system.

According to the inclusion and exclusion criteria, the data of 1 475 patients who underwent ultrasound-guided CNB for breast masses at the Baoji Central Hospital from January 2024 to December 2025 were enrolled in this prospective study. Touch imprints were simultaneously prepared from biopsy specimens for ROSE. ROSE interpretation was performed with the Yokohama reporting system by the International Academy of Cytology as the classification standard, and the results were divided into benign lesion, atypical lesion, suspicious for malignancy and malignant tumour. Using histological diagnosis as the gold standard, the concordance between ROSE classification results based on the Yokohama reporting system and histological diagnosis was compared and analysed. The Kappa consistency test was adopted to evaluate the diagnostic efficacy of ROSE for breast masses, and patient triage was guided according to the classification results.

Among the 1 475 specimens, ROSE classifications based on the Yokohama reporting system showed 770 benign lesions, 74 atypical lesions, 43 suspicious malignancies and 588 malignant tumours. The diagnostic accuracy compared with histological diagnosis was 99.5% (766/770), 55.4% (41/74), 93.0% (40/43) and 100% (588/588), respectively. The sensitivity, specificity and overall accuracy of ROSE classification based on the Yokohama reporting system against the histopathological gold standard were 94.58% (628/664), 99.63% (808/811) and 97.36% (1 436/1 475), respectively. The positive predictive value and negative predictive value were 99.52% (628/631) and 95.73% (808/844). The ROSE classification of breast masses had good consistency with histological diagnosis (κ=0.946, P<0.05). All 770 patients with ROSE classification of benign lesion received day-case surgery. Intraoperative frozen pathology revealed 3 cases of malignancy and 1 case of desmoid fibromatosis, for which extended surgery was subsequently performed. The remaining 766 patients consistent with ROSE results were discharged on the next day. Patients with ROSE results of atypical lesion, suspicious malignancy and malignant tumour waited for the final histological report on the following day. Patients with histologically confirmed benign lesions underwent minimally invasive surgery; patients with borderline tumours received extended surgery; patients with malignant tumours received neoadjuvant therapy, modified radical mastectomy or breast-conserving surgery.

The Yokohama reporting system can be applied to ROSE of touch imprints from CNB of breast masses. The categories of benign lesion, suspicious malignancy and malignant tumour demonstrate high concordance with histological diagnosis. This system exhibits high diagnostic efficacy in evaluating the nature of breast masses and helps optimise the diagnosis and treatment workflow for patients with breast masses.