Abstract P5-05-01: Prevalance of crown-like structures of the breast, a histologic biomarker linked to obesity: A retrospective study of 99 cases

Abstract

Abstract Introduction: Breast cancer risk is multifactorial, and depends partly on obesity and related metabolic imbalances, including inflammation. Obesity is increasing worldwide, and is a known cancer risk albeit with complex mechanisms. Previous reports (Morris et al., 2011; Iyengar et al., 2015) indicate that local inflammation can be seen histologically as a rings of macrophages around necrotic adipocytes ("crown-like structures of the breast", CLS-B). Our goal was to determine the prevalence of CLS-B in routine specimens from a cohort of patients with known BMI. Methods: We retrieved archival H&E slides from a breast cancer cohort (N=99) previously characterized for BMI and fasting plasma/serum metabolic factors. Two pathologists reviewed all available sections of white adipose tissue not adjacent to tumour (median 7 blocks/case), excluding fat necrosis and mastitis, blinded to correlative data/BMI. We recorded the presence/absence and numbers of CLS-B, defined as a continuous ring of macrophages surrounding an adipocyte. Paraffin blocks were available in a subset (N=72) and a representative block was immunostained for CD68 to highlight CLS-B. For all cases, the average fat vacuole size was determined by digital image analysis (NIH ImageJ Software). We performed correlative statistics between CLS-B status and clinical data (χ2, Wilcoxon rank-sum tests). Results: CLS-B were present in 37 of 99 cases (37%). When present the total number of CLS-B ranged from 1 to 18 (mean=4.3, median=3). CLS-B were detected in 7/10 (70%) patients with BMI >30 vs. 30/89 (34%) with BMI ≤ 30 (p=0.02). CLS-B also trended to higher prevalence in women over 60 compared to women under 60 (12/20, 60% vs. 25/79, 32%, p = 0.063). There was no significant association of CLS-B status with tumor T- and N-stage or grade (all P>0.4). The median C-reactive protein in the group with CLS-B was 1.5 mg/L vs. 0.8 mg/L in the group without CLS-B (P=0.10) There was no significant association of CLS-B with insulin, glucose, HOMA, leptin, adiponectin, total cholesterol, triglycerides, HDL cholesterol, LDL cholesterol, or IGF-1 (all P>0.27). The average fat globule area determined by image analysis correlated significantly with BMI (Spearman correlation 0.54, p<0.0001) but not to the presence of CLS-B (p=0.102). Within the subset immunostained for CD68, 32/72 (44%) had CLS-B on the original H&E sections, whereas 13/72 (18%) had CLS-B on the representative CD68-stained section. This corresponded to a false negative in 22/59 (37%) CD68-negative cases, and increased detection in 3/13 of the CD68-positive cases. Conclusion: In our cohort, obesity is correlated with elevated tissue inflammation as seen by the presence of CLS-B, but CLS-B is not correlated with metabolic markers. CLS-B are well appreciated on routine H&E sections; however, more work is needed to find a practical approach to both ancillary testing (e.g. CD68) and quantitation. Our work independently confirms the association of CLS-B with obesity, and supports the concept that CLS-B is a tissue biomarker of obesity-related inflammation. (Z.E. was co-principal author.) Citation Format: Chang MC, Eslami Z, Ennis M, Goodwin PJ. Prevalance of crown-like structures of the breast, a histologic biomarker linked to obesity: A retrospective study of 99 cases. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P5-05-01.