Evaluating the effectiveness of pre-operative diagnosis of ovarian cancer using minimally invasive liquid biopsies by combining serum human epididymis protein 4 and cell-free DNA in patients with an ovarian mass.

OBJECTIVE: To assess the feasibility of scalable, objective, and minimally invasive liquid biopsy-derived biomarkers such as cell-free DNA copy number profiles, human epididymis protein 4 (HE4), and cancer antigen 125 (CA125) for pre-operative risk assessment of early-stage ovarian cancer in a clinically representative and diagnostically challenging population and to compare the performance of these biomarkers with the Risk of Malignancy Index (RMI). METHODS: In this case-control study, we included 100 patients with an ovarian mass clinically suspected to be early-stage ovarian cancer. Of these 100 patients, 50 were confirmed to have a malignant mass (cases) and 50 had a benign mass (controls). Using WisecondorX, an algorithm used extensively in non-invasive prenatal testing, we calculated the benign-calibrated copy number profile abnormality score. This score represents how different a sample is from benign controls based on copy number profiles. We combined this score with HE4 serum concentration to separate cases and controls. RESULTS: Combining the benign-calibrated copy number profile abnormality score with HE4, we obtained a model with a significantly higher sensitivity (42% vs 0%; p<0.002) at 99% specificity as compared with the RMI that is currently employed in clinical practice. Investigating performance in subgroups, we observed especially large differences in the advanced stage and non-high-grade serous ovarian cancer groups. CONCLUSION: This study demonstrates that cell-free DNA can be successfully employed to perform pre-operative risk of malignancy assessment for ovarian masses; however, results warrant validation in a more extensive clinical study.

Imprint of parity and age at first pregnancy on the genomic landscape of subsequent breast cancer.

BACKGROUND: Although parity and age at first pregnancy are among the most known extrinsic factors that modulate breast cancer risk, their impact on the biology of subsequent breast cancer has never been explored in depth. Recent data suggest that pregnancy-induced tumor protection is different according to breast cancer subtypes, with parity and young age at first pregnancy being associated with a marked reduction in the risk of developing luminal subtype but not triple negative breast cancer. In this study, we investigated the imprint of parity and age at first pregnancy on the pattern of somatic mutations, somatic copy number alterations, transcriptomic profiles, and tumor immune microenvironment by assessing tumor-infiltrating lymphocytes (TILs) levels of subsequent breast cancer. METHODS: A total of 313 patients with primary breast cancer with available whole genome, RNA sequencing, and TILs data were included in this study. We used a multivariate analysis adjusted for age at diagnosis, pathological stage, molecular subtypes, and histological subtypes. We compared nulliparous vs. parous, late parous vs. early parous, and nulliparous vs. pregnancy-associated breast cancer (PABC) patients. Late and early parous patients were grouped by using the median age at first pregnancy. PABC was defined as patients diagnosed up to 10 years postpartum. RESULTS: Genomic alterations of breast cancer were associated with age at first pregnancy but not with parity status alone. Independently of clinicopathological features, early parous patients developed tumors characterized by a higher number of Indels (Padj = 0.002), a lower frequency of CDH1 mutations (1.2% vs. 12.7%; Padj = 0.013), a higher frequency of TP53 mutations (50% vs. 22.5%; Padj = 0.010), and MYC amplification (28% vs. 7%; Padj = 0.008). PABC were associated with increased TILs infiltration (Padj = 0.0495). CONCLUSIONS: These findings highlight an unprecedented link between reproductive history and the genomic landscape of subsequent breast cancer. We further hypothesize that TP53-mutant premalignant lesions could be less susceptible to the protective effect of an early parity, which might explain the difference of parity-induced protection according to breast cancer subtypes. This work also advocates that reproductive history should be routinely collected in future large-scale genomic studies addressing the biology of female cancers.

PatchSorter: a high throughput deep learning digital pathology tool for object labeling.

The discovery of patterns associated with diagnosis, prognosis, and therapy response in digital pathology images often requires intractable labeling of large quantities of histological objects. Here we release an open-source labeling tool, PatchSorter, which integrates deep learning with an intuitive web interface. Using >100,000 objects, we demonstrate a >7x improvement in labels per second over unaided labeling, with minimal impact on labeling accuracy, thus enabling high-throughput labeling of large datasets.

HRD-related morphology discovery in breast cancer by controlling for confounding factors.

Lazard et al.1 predict homologous recombination deficiency from hematoxylin and eosin-stained slides of breast cancer tissue using deep learning. By controlling for technical artifacts on a curated dataset, the model puts forward novel HRD-related morphologies in luminal breast cancers.

DeepSMILE: Contrastive self-supervised pre-training benefits MSI and HRD classification directly from H&E whole-slide images in colorectal and breast cancer.

We propose a Deep learning-based weak label learning method for analyzing whole slide images (WSIs) of Hematoxylin and Eosin (H&E) stained tumor tissue not requiring pixel-level or tile-level annotations using Self-supervised pre-training and heterogeneity-aware deep Multiple Instance LEarning (DeepSMILE). We apply DeepSMILE to the task of Homologous recombination deficiency (HRD) and microsatellite instability (MSI) prediction. We utilize contrastive self-supervised learning to pre-train a feature extractor on histopathology tiles of cancer tissue. Additionally, we use variability-aware deep multiple instance learning to learn the tile feature aggregation function while modeling tumor heterogeneity. For MSI prediction in a tumor-annotated and color normalized subset of TCGA-CRC (n=360 patients), contrastive self-supervised learning improves the tile supervision baseline from 0.77 to 0.87 AUROC, on par with our proposed DeepSMILE method. On TCGA-BC (n=1041 patients) without any manual annotations, DeepSMILE improves HRD classification performance from 0.77 to 0.81 AUROC compared to tile supervision with either a self-supervised or ImageNet pre-trained feature extractor. Our proposed methods reach the baseline performance using only 40% of the labeled data on both datasets. These improvements suggest we can use standard self-supervised learning techniques combined with multiple instance learning in the histopathology domain to improve genomic label classification performance with fewer labeled data.

Tumour-infiltrating lymphocytes (TILs) and BRCA-like status in stage III breast cancer patients randomised to adjuvant intensified platinum-based chemotherapy versus conventional chemotherapy.

BACKGROUND: The prognostic value of tumour-infiltrating lymphocytes (TILs) differs by breast cancer (BC) subtype. The aim of this study was to evaluate TILs in stage III BC in the context of BRCA1/2-like phenotypes and association with outcome and benefit of intensified platinum-based chemotherapy. PATIENTS AND METHODS: Patients participated in a randomised controlled trial of adjuvant intensified platinum-based chemotherapy versus conventional anthracycline-based chemotherapy carried out between 1993 and 1999 in stage III BC. Stromal TILs were scored according to International guidelines in these human epidermal growth factor receptor 2 (HER2)-negative tumours. BRCA-profiles were determined using Comparative Genomic Hybridization. RESULTS: TIL levels were evaluated in 248 BCs. High TILs were associated with Triple Negative BC (TNBC). BRCA-like tumours harboured higher TILs compared to non-BRCA-like tumours (median TILs of 20% versus 10%, p < 0.01). TIL levels in BRCA1-like tumours were higher compared to BRCA2-like tumours (median TILs of 20% versus 10%, p < 0.001). These correlations remained significant within the oestrogen (ER)-positive subgroup, however not within the TNBC subgroup. In this stage III BC cohort, high TIL level was associated with favourable outcome (TILs per 10% increment, recurrence-free survival (RFS): multivariate hazard ratio (HR) 0.82, 95% confidence interval (CI) 0.71-0.94, p = 0.01; overall survival (OS): multivariate HR 0.80, 95% CI 0.68-0.94, p = 0.01). There was no significant interaction between TILs and benefit of intensified platinum-based chemotherapy. CONCLUSION: In this high-risk breast cancer cohort, high TILs were associated with TNBC and BRCA1-like status. Within the ER-positive subgroup, TIL levels were higher in BRCA1-like compared to BRCA2-like tumours. When adjusted for clinical characteristics, TILs were significantly associated with a more favourable outcome in stage III BC patients.