Allison Hall

Despite increasing evidence supporting the clinical relevance of tumour infiltrating lymphocytes (TILs) in invasive breast cancer, TIL spatial variability within ductal carcinoma in situ (DCIS) samples and its association with progression are not well understood. To characterise tissue spatial architecture and the microenvironment of DCIS, we designed and validated a new deep learning pipeline, UNMaSk. Following automated detection of individual DCIS ducts using a new method IM-Net, we applied spatial tessellation to create virtual boundaries for each duct. To study local TIL infiltration for each duct, DRDIN was developed for mapping the distribution of TILs. In a dataset comprising grade 2-3 pure DCIS and DCIS adjacent to invasive cancer (adjacent DCIS), we found that pure DCIS cases had more TILs compared to adjacent DCIS. However, the colocalisation of TILs with DCIS ducts was significantly lower in pure DCIS compared to adjacent DCIS, which may suggest a more inflamed tissue ecology local to DCIS ducts in adjacent DCIS cases. Our study demonstrates that technological developments in deep convolutional neural networks and digital pathology can enable an automated morphological and microenvironmental analysis of DCIS, providing a new way to study differential immune ecology for individual ducts and identify new markers of progression.

INTRODUCTION: Breast cancer is a major cause of morbidity and mortality for women in Sub-Saharan Africa and for black American women. There is evidence that the pathologic characteristics of breast cancers in both African women and black American women may differ from their counterparts of European ancestry. However, despite the great burden of disease, data on pathologic features of breast carcinoma in Sub-Saharan Africa is limited and often contradictory. This lack of information makes it difficult to prioritize resource use in efforts to improve breast cancer outcomes in the region. METHODS: We examined consecutive cases of breast cancer in Tanzanian women (n = 83), black American women (n = 120), and white American women (n = 120). Each case was assessed for tumor type, grade, mitotic count, ER and HER2 status, and tumor infiltrating lymphocyte involvement. RESULTS: The Tanzanian subjects were younger and had higher stage tumors than the subjects in either American group. Breast cancers in the Tanzanian and black American groups were more likely to be high grade (p = 0.008), to have a high mitotic rate (p<0.0001), and to be ER-negative (p<0.001) than the tumors in the white American group. Higher levels of tumor infiltrating lymphocyte involvement were seen among Tanzanian and black American subjects compared to white American subjects (p = 0.0001). Among all subjects, tumor infiltrating lymphocyte levels were higher in tumors with a high mitotic rate. Among Tanzanian and black American subjects, tumor infiltrating lymphocyte levels were higher in ER-negative tumors. These findings have implications for treatment priorities for breast cancer in Tanzania and other Sub-Saharan African countries.

The study evaluated morphologic patterns, mutational profiles, and β-catenin immunohistochemistry (IHC) in copy-number low (CNL) endometrial adenocarcinomas (EAs). CNL EAs (n=19) with next-generation or whole genome sequencing results and available tissue for IHC were identified from our institutional database. Clinical data and histologic slides were reviewed. IHC for β-catenin was performed and correlated with mutation status. Images of digital slides of CNL EAs from The Cancer Genome Atlas (TCGA) database (n=90) were blindly reviewed by 4 pathologists, and morphology was correlated with mutation status. Categorical variables were analyzed using the Fisher exact test, and agreement was assessed using Fleiss κ. CTNNB1 mutations were present in 63% (12/19) of CNL EAs. β-catenin nuclear localization was present in 83% of CTNNB1-mutated tumors (10/12) and in 0% (0/7) of CTNNB1-wildtype tumors (sensitivity 0.83, specificity 1.00). Squamous differentiation (SD) was present in 47% (9/19) and was more often observed in CTNNB1-mutated tumors (P=0.02). Mucinous differentiation (MD) was associated with KRAS mutations (P<0.01). Digital image review of TCGA CNL EAs revealed that pathologist agreement on SD was strong (κ=0.82), whereas agreement on MD was weak (κ=0.48). Pathologists identified SD in 22% (20/90), which was significantly associated with the presence of CTNNB1 mutations (P<0.01). CNL EAs demonstrate several morphologies with divergent molecular profiles. SD was significantly associated with CTNNB1 mutations and nuclear localization of β-catenin in these tumors. Nuclear expression of β-catenin is a sensitive and specific IHC marker for CTNNB1 mutations in CNL EAs. CNL EAs with KRAS mutations often displayed MD.

OBJECTIVES: This study assessed the prevalence, general interest, and barriers to implementing global health curricula in pathology residency programs. METHODS: We conducted a survey of 166 US pathology residency programs. RESULTS: Thirty-two (195) of 166 programs responded. Of these, 13% have a formalized global health program (n = 4), and the majority indicated at least some general interest in global health among trainees (88%, n = 28) and faculty (94%, n = 30), albeit at a low to moderate level. Funding limitations, regulatory constraints, and insufficient knowledge of global health were frequently cited barriers to developing a global health program. CONCLUSIONS: Few US pathology departments incorporate global health education into postgraduate training. The importance of pathology in global health has been underappreciated, despite its critical role in the delivery of health care in resource-limited settings. One solution is for pathology departments to expand global health educational opportunities for trainees.