Benign and borderline phyllodes tumors of the breast: Clinicopathologic analysis of 205 cases with emphasis on the surgical margin status and local recurrence rate.

BACKGROUND: The management of benign and borderline phyllodes tumors of the breast with a positive surgical margin is still controversial. Our aim in this study was to evaluate the impact of surgical margin status on the local recurrence rate of benign and borderline phyllodes tumors. METHODS: We reviewed 205 phyllodes tumors (191 benign, 14 borderline) that were surgically excised at our hospital between 2005 and 2019. Follow-up information extending to at least 6 months after surgery was retrieved from the clinical, radiology, and pathology records. RESULTS: The initial surgical margin was negative in 54 (26%) cases, close (≤ 1 mm) in 29 (14%) cases, and positive in 122 (60%) cases. Approximately half of the cases with a close margin and two-third of the cases with a positive margin underwent re-excision to obtain negative margins. Three (2.3%) local recurrences were observed among 131 cases with follow-up information, all three with benign phyllodes tumor. Of these three patients, one had a positive final margin, and two had negative final margins. There was no significant difference in the rate of local recurrence between PT with a positive surgical margin versus a close and negative margin. CONCLUSION: The study results suggest that close clinical and radiologic follow-up may provide a better course of management rather than re-excision when managing positive margins in benign and borderline phyllodes tumors.

High EZH2 expression in ductal carcinoma in situ diagnosed on breast core needle biopsy is an independent predictive factor for upgrade on surgical excision.

PURPOSE: Approximately 25 % of DCIS diagnosed on breast core needle biopsy (CNB) is upgraded to invasive carcinoma on surgical excision. Risk factors to predict the upgrade on excision are not well established, leading many patients to be over or under-treated. EZH2 was shown to be associated with aggressive behavior of cancer from many sites, including breast cancer. We aimed to analyze EZH2 expression and tumor infiltrating lymphocytes (TILs) in DCIS as predictive factors for an upgrade on excision. METHODS: We assessed EZH2 expression in 34 DCIS cases diagnosed on CNB and upgraded to invasive carcinoma on excision. Then, we compared these cases with 60 control cases that were not upgraded on excision. A staining score for DCIS (0-12) was obtained by multiplying the staining intensity (0-3) and the percentage of positive cells (1-4). The nuclear staining score ≥6 was considered as 'high' expression. RESULTS: 46 of 94 (49 %) DCIS on CNB showed high EZH2 expression. EZH2 expression was directly correlated with TILs density, nuclear grade, HER2 expression, Ki-67 index and negative ER status. On univariate analysis, upgrade on excision was associated with high EZH2 expression, high TILs density, negative ER status and high Ki-67 index. Multivariate analysis revealed the high EZH2 expression as the only independent predictive factor for upgrade on excision. CONCLUSIONS: Our study revealed the high EZH2 expression as the only independent predictive factor for an upgrade on excision. Future studies should focus on the evaluation of EZH2 expression in tumor-microenvironment interaction in terms of diagnostic, treatment and prognostic purposes.

Harnessing oncolytic virus-mediated antitumor immunity in an infected cell vaccine.

Treatment of permissive tumors with the oncolytic virus (OV) VSV-Δ51 leads to a robust antitumor T-cell response, which contributes to efficacy; however, many tumors are not permissive to in vivo treatment with VSV-Δ51. In an attempt to channel the immune stimulatory properties of VSV-Δ51 and broaden the scope of tumors that can be treated by an OV, we have developed a potent oncolytic vaccine platform, consisting of tumor cells infected with VSV-Δ51. We demonstrate that prophylactic immunization with this infected cell vaccine (ICV) protected mice from subsequent tumor challenge, and expression of granulocyte-monocyte colony stimulating factor (GM-CSF) by the virus (VSVgm-ICV) increased efficacy. Immunization with VSVgm-ICV in the VSV-resistant B16-F10 model induced maturation of dendritic and natural killer (NK) cell populations. The challenge tumor is rapidly infiltrated by a large number of interferon γ (IFNγ)-producing T and NK cells. Finally, we demonstrate that this approach is robust enough to control the growth of established tumors. This strategy is broadly applicable because of VSV's extremely broad tropism, allowing nearly all cell types to be infected at high multiplicities of infection in vitro, where the virus replication kinetics outpace the cellular IFN response. It is also personalized to the unique tumor antigen(s) displayed by the cancer cell.

Visual discrimination learning in zebrafish (Danio rerio).

Three experiments demonstrated visual discrimination learning in zebrafish (Danio rerio). In each experiment, zebrafish were given a choice between two visually distinct arms of a T-maze. Choice of one stimulus was always followed by a food reward, but choice of the other stimulus was not rewarded. Different colored sleeves fitted around the arms of the T-maze were used in Experiments 1 (green and purple) and 2 (red and blue). The stimuli used in Experiment 3 were white sleeves lined with horizontal or vertical black stripes. In all three experiments, zebrafish acquired a significant preference for the stimulus that led to a food reward. Experiments 1 and 2 also showed that zebrafish could learn a reversal of the discrimination. Finally, the effect of discontinuing food rewards was examined after reversal training in Experiment 2 and after initial discrimination training in Experiments 1 and 3. Non-reinforcement led to a decrease in correct responding in Experiments 2 and 3 independent of stimulus identity, but to an asymmetrical pattern of responding in Experiment 1. The median latency to make a choice response decreased over the course of acquisition in all three experiments; during extinction, median response times did not change at all in Experiment 1 and increased only very slightly in Experiment 2, but showed a substantial increase in Experiment 3. The implications of these results for the zebrafish as a model system for genetic studies of learning and memory are discussed.