Uninformed Teacher-Student for hard-samples distillation in weakly supervised mitosis localization.

BACKGROUND AND OBJECTIVE: Mitotic activity is a crucial biomarker for diagnosing and predicting outcomes for different types of cancers, particularly breast cancer. However, manual mitosis counting is challenging and time-consuming for pathologists, with moderate reproducibility due to biopsy slide size, low mitotic cell density, and pattern heterogeneity. In recent years, deep learning methods based on convolutional neural networks (CNNs) have been proposed to address these limitations. Nonetheless, these methods have been hampered by the available data labels, which usually consist only of the centroids of mitosis, and by the incoming noise from annotated hard negatives. As a result, complex algorithms with multiple stages are often required to refine the labels at the pixel level and reduce the number of false positives. METHODS: This article presents a novel weakly supervised approach for mitosis detection that utilizes only image-level labels on histological hematoxylin and eosin (H&E) images, avoiding the need for complex labeling scenarios. Also, an Uninformed Teacher-Student (UTS) pipeline is introduced to detect and distill hard samples by comparing weakly supervised localizations and the annotated centroids, using strong augmentations to enhance uncertainty. Additionally, an automatic proliferation score is proposed that mimicks the pathologist-annotated mitotic activity index (MAI). The proposed approach is evaluated on three publicly available datasets for mitosis detection on breast histology samples, and two datasets for mitotic activity counting in whole-slide images. RESULTS: The proposed framework achieves competitive performance with relevant prior literature in all the datasets used for evaluation without explicitly using the mitosis location information during training. This approach challenges previous methods that rely on strong mitosis location information and multiple stages to refine false positives. Furthermore, the proposed pipeline for hard-sample distillation demonstrates promising dataset-specific improvements. Concretely, when the annotation has not been thoroughly refined by multiple pathologists, the UTS model offers improvements of up to ∼4% in mitosis localization, thanks to the detection and distillation of uncertain cases. Concerning the mitosis counting task, the proposed automatic proliferation score shows a moderate positive correlation with the MAI annotated by pathologists at the biopsy level on two external datasets. CONCLUSIONS: The proposed Uninformed Teacher-Student pipeline leverages strong augmentations to distill uncertain samples and measure dissimilarities between predicted and annotated mitosis. Results demonstrate the feasibility of the weakly supervised approach and highlight its potential as an objective evaluation tool for tumor proliferation.

Digital image analysis and machine learning-assisted prediction of neoadjuvant chemotherapy response in triple-negative breast cancer.

BACKGROUND: Pathological complete response (pCR) is associated with favorable prognosis in patients with triple-negative breast cancer (TNBC). However, only 30-40% of TNBC patients treated with neoadjuvant chemotherapy (NAC) show pCR, while the remaining 60-70% show residual disease (RD). The role of the tumor microenvironment in NAC response in patients with TNBC remains unclear. In this study, we developed a machine learning-based two-step pipeline to distinguish between various histological components in hematoxylin and eosin (H&E)-stained whole slide images (WSIs) of TNBC tissue biopsies and to identify histological features that can predict NAC response. METHODS: H&E-stained WSIs of treatment-naïve biopsies from 85 patients (51 with pCR and 34 with RD) of the model development cohort and 79 patients (41 with pCR and 38 with RD) of the validation cohort were separated through a stratified eightfold cross-validation strategy for the first step and leave-one-out cross-validation strategy for the second step. A tile-level histology label prediction pipeline and four machine-learning classifiers were used to analyze 468,043 tiles of WSIs. The best-trained classifier used 55 texture features from each tile to produce a probability profile during testing. The predicted histology classes were used to generate a histology classification map of the spatial distributions of different tissue regions. A patient-level NAC response prediction pipeline was trained with features derived from paired histology classification maps. The top graph-based features capturing the relevant spatial information across the different histological classes were provided to the radial basis function kernel support vector machine (rbfSVM) classifier for NAC treatment response prediction. RESULTS: The tile-level prediction pipeline achieved 86.72% accuracy for histology class classification, while the patient-level pipeline achieved 83.53% NAC response (pCR vs. RD) prediction accuracy of the model development cohort. The model was validated with an independent cohort with tile histology validation accuracy of 83.59% and NAC prediction accuracy of 81.01%. The histological class pairs with the strongest NAC response predictive ability were tumor and tumor tumor-infiltrating lymphocytes for pCR and microvessel density and polyploid giant cancer cells for RD. CONCLUSION: Our machine learning pipeline can robustly identify clinically relevant histological classes that predict NAC response in TNBC patients and may help guide patient selection for NAC treatment.

Proliferation and immunohistochemistry for p53, CD25 and CK20 in predicting prognosis of non-muscle invasive papillary urothelial carcinomas.

Non-muscle invasive papillary urothelial carcinoma is a prevalent disease with a high recurrence tendency. Good prognostic and reproducible biomarkers for tumor recurrence and disease progression are lacking. Currently, WHO grade and tumor stage are essential in risk stratification and treatment decision-making. Here we present the prognostic value of proliferation markers (Ki67, mitotic activity index (MAI) and PPH3) together with p53, CD25 and CK20 immunohistochemistry (IHC). In this population-based retrospective study, 349 primary non-muscle invasive bladder cancers (NMIBC) were available. MAI and PPH3 were calculated manually according to highly standardized previously described methods, Ki-67 by the semi-automated QPRODIT quantification system, p53 and CD25 by the fully automated digital image analysis program Visipharm® and CK20 with the help of the semi-quantitative immunoreactive score (IRS). Survival analyses with log rank test, as well as univariate and multivariate Cox regression analyses were performed for all investigated variables. Age and multifocality were the only significant variables for tumor recurrence. All investigated variables, except gender, were significantly associated with stage progression. In multivariate analysis, MAI was the only prognostic variable for stage progression (p<0.001).

Prenatal BRCA1 epimutations contribute significantly to triple-negative breast cancer development.

BACKGROUND: Normal cell BRCA1 epimutations have been associated with increased risk of triple-negative breast cancer (TNBC). However, the fraction of TNBCs that may have BRCA1 epimutations as their underlying cause is unknown. Neither are the time of occurrence and the potential inheritance patterns of BRCA1 epimutations established. METHODS: To address these questions, we analyzed BRCA1 methylation status in breast cancer tissue and matched white blood cells (WBC) from 408 patients with 411 primary breast cancers, including 66 TNBCs, applying a highly sensitive sequencing assay, allowing allele-resolved methylation assessment. Furthermore, to assess the time of origin and the characteristics of normal cell BRCA1 methylation, we analyzed umbilical cord blood of 1260 newborn girls and 200 newborn boys. Finally, we assessed BRCA1 methylation status among 575 mothers and 531 fathers of girls with (n = 102) and without (n = 473) BRCA1 methylation. RESULTS: We found concordant tumor and mosaic WBC BRCA1 epimutations in 10 out of 66 patients with TNBC and in four out of six patients with estrogen receptor (ER)-low expression (< 10%) tumors (combined: 14 out of 72; 19.4%; 95% CI 11.1-30.5). In contrast, we found concordant WBC and tumor methylation in only three out of 220 patients with 221 ER ≥ 10% tumors and zero out of 114 patients with 116 HER2-positive tumors. Intraindividually, BRCA1 epimutations affected the same allele in normal and tumor cells. Assessing BRCA1 methylation in umbilical WBCs from girls, we found mosaic, predominantly monoallelic BRCA1 epimutations, with qualitative features similar to those in adults, in 113/1260 (9.0%) of individuals, but no correlation to BRCA1 methylation status either in mothers or fathers. A significantly lower fraction of newborn boys carried BRCA1 methylation (9/200; 4.5%) as compared to girls (p = 0.038). Similarly, WBC BRCA1 methylation was found less common among fathers (16/531; 3.0%), as compared to mothers (46/575; 8.0%; p = 0.0003). CONCLUSIONS: Our findings suggest prenatal BRCA1 epimutations might be the underlying cause of around 20% of TNBC and low-ER expression breast cancers. Such constitutional mosaic BRCA1 methylation likely arise through gender-related mechanisms in utero, independent of Mendelian inheritance.

Homologous Recombination Deficiency Across Subtypes of Primary Breast Cancer.

PURPOSE: Homologous recombination deficiency (HRD) is highly prevalent in triple-negative breast cancer (TNBC) and associated with response to PARP inhibition (PARPi). Here, we studied the prevalence of HRD in non-TNBC to assess the potential for PARPi in a wider group of patients with breast cancer. METHODS: HRD status was established using targeted gene panel sequencing (360 genes) and BRCA1 methylation analysis of pretreatment biopsies from 201 patients with primary breast cancer in the phase II PETREMAC trial (ClinicalTrials.gov identifier: NCT02624973). HRD was defined as mutations in BRCA1, BRCA2, BRIP1, BARD1, or PALB2 and/or promoter methylation of BRCA1 (strict definition; HRD-S). In secondary analyses, a wider definition (HRD-W) was used, examining mutations in 20 additional genes. Furthermore, tumor BRCAness (multiplex ligation-dependent probe amplification), PAM50 subtyping, RAD51 nuclear foci to test functional HRD, tumor-infiltrating lymphocyte (TIL), and PD-L1 analyses were performed. RESULTS: HRD-S was present in 5% of non-TNBC cases (n = 9 of 169), contrasting 47% of the TNBC tumors (n = 15 of 32). HRD-W was observed in 23% of non-TNBC (n = 39 of 169) and 59% of TNBC cases (n = 19 of 32). Of 58 non-TNBC and 30 TNBC biopsies examined for RAD51 foci, 4 of 4 (100%) non-TNBC and 13 of 14 (93%) TNBC cases classified as HRD-S had RAD51 low scores. In contrast, 4 of 17 (24%) non-TNBC and 15 of 19 (79%) TNBC biopsies classified as HRD-W exhibited RAD51 low scores. Of nine non-TNBC tumors with HRD-S status, only one had a basal-like PAM50 signature. There was a high concordance between HRD-S and either BRCAness, high TIL density, or high PD-L1 expression (each P < .001). CONCLUSION: The prevalence of HRD in non-TNBC suggests that therapy targeting HRD should be evaluated in a wider breast cancer patient population. Strict HRD criteria should be implemented to increase diagnostic precision with respect to functional HRD.

Digital image analysis and machine learning-assisted prediction of neoadjuvant chemotherapy response in triple-negative breast cancer.

Background: Pathological complete response (pCR) is associated with favorable prognosis in patients with triple-negative breast cancer (TNBC). However, only 30-40% of TNBC patients treated with neoadjuvant chemotherapy (NAC) show pCR, while the remaining 60-70% show residual disease (RD). The role of the tumor microenvironment (TME) in NAC response in patients with TNBC remains unclear. In this study, we developed a machine learning-based two-step pipeline to distinguish between various histological components in hematoxylin and eosin (H&E)-stained whole slide images (WSIs) of TNBC tissue biopsies and to identify histological features that can predict NAC response. Methods: H&E-stained WSIs of treatment-naïve biopsies from 85 patients (51 with pCR and 34 with RD) were separated through a stratified 8-fold cross validation strategy for the first step and leave one out cross validation strategy for the second step. A tile-level histology label prediction pipeline and four machine learning classifiers were used to analyze 468,043 tiles of WSIs. The best-trained classifier used 55 texture features from each tile to produce a probability profile during testing. The predicted histology classes were used to generate a histology classification map of the spatial distributions of different tissue regions. A patient-level NAC response prediction pipeline was trained with features derived from paired histology classification maps. The top graph-based features capturing the relevant spatial information across the different histological classes were provided to the radial basis function kernel support vector machine (rbfSVM) classifier for NAC treatment response prediction. Results: The tile-level prediction pipeline achieved 86.72% accuracy for histology class classification, while the patient-level pipeline achieved 83.53% NAC response (pCR vs. RD) prediction accuracy. The histological class pairs with the strongest NAC response predictive ability were tumor and tumor tumor-infiltrating lymphocytes for pCR and microvessel density and polyploid giant cancer cells for RD. Conclusion: Our machine learning pipeline can robustly identify clinically relevant histological classes that predict NAC response in TNBC patients and may help guide patient selection for NAC treatment.

Importance of endocrine treatment adherence and persistence in breast cancer survivorship: a systematic review.

PURPOSE: Adjuvant endocrine treatment is essential for treating luminal subtypes of breast cancer, which constitute 75% of all breast malignancies. However, the detrimental side effects of treatment make it difficult for many patients to complete the guideline-required treatment. Such non-adherence may jeopardize the lifesaving ability of anti-estrogen therapy. In this systematic review, we aimed to assess the consequences of non-adherence and non-persistence from available studies meeting strict statistical and clinical criteria. METHODS: A systematic literature search was performed using several databases, yielding identification of 2,026 studies. After strict selection, 14 studies were eligible for systematic review. The review included studies that examined endocrine treatment non-adherence (patients not taking treatment as prescribed) or non-persistence (patients stopping treatment prematurely), in terms of the effects on event-free survival or overall survival among women with non-metastatic breast cancer. RESULTS: We identified 10 studies measuring the effects of endocrine treatment non-adherence and non-persistence on event-free survival. Of these studies, seven showed significantly poorer survival for the non-adherent or non-persistent patient groups, with hazard ratios (HRs) ranging from 1.39 (95% CI, 1.07 to 1.53) to 2.44 (95% CI, 1.89 to 3.14). We identified nine studies measuring the effects of endocrine treatment non-adherence and non-persistence on overall survival. Of these studies, seven demonstrated significantly reduced overall survival in the groups with non-adherence and non-persistence, with HRs ranging from 1.26 (95% CI, 1.11 to 1.43) to 2.18 (95% CI, 1.99 to 2.39). CONCLUSION: The present systematic review demonstrates that non-adherence and non-persistence to endocrine treatment negatively affect event-free and overall survival. Improved follow-up, with focus on adherence and persistence, is vital for improving health outcomes among patients with non-metastatic breast cancer.

Predicting neoadjuvant treatment response in triple-negative breast cancer using machine learning.

Background: Neoadjuvant chemotherapy (NAC) is the standard treatment for early-stage triple negative breast cancer (TNBC). The primary endpoint of NAC is a pathological complete response (pCR). NAC results in pCR in only 30%â€"40% of TNBC patients. Tumor-infiltrating lymphocytes (TILs), Ki67 and phosphohistone H3 (pH3) are a few known biomarkers to predict NAC response. Currently, systematic evaluation of the combined value of these biomarkers in predicting NAC response is lacking. In this study, the predictive value of markers derived from H&E and IHC stained biopsy tissue was comprehensively evaluated using a supervised machine learning (ML)-based approach. Identifying predictive biomarkers could help guide therapeutic decisions by enabling precise stratification of TNBC patients into responders and partial or non-responders. Methods: Serial sections from core needle biopsies (n=76) were stained with H&E, and immunohistochemically for the Ki67 and pH3 markers, followed by whole slide image (WSI) generation. The resulting WSI triplets were co-registered with H&E WSIs serving as the reference. Separate mask region-based CNN (MRCNN) models were trained with annotated H&E, Ki67 and pH3 images for detecting tumor cells, stromal and intratumoral TILs (sTILs and tTILs), Ki67 + , and pH3 + cells. Top image patches with a high density of cells of interest were identified as hotspots. Best classifiers for NAC response prediction were identified by training multiple ML models, and evaluating their performance by accuracy, area under curve, and confusion matrix analyses. Results: Highest prediction accuracy was achieved when hotspot regions were identified by tTIL counts and each hotspot was represented by measures of tTILs, sTILs, tumor cells, Ki67 + , and pH3 + features. Regardless of the hotspot selection metric, a complementary use of multiple histological features (tTILs, sTILs) and molecular biomarkers (Ki67 and pH3) resulted in top ranked performance at the patient level. Conclusions: Overall, our results emphasize that prediction models for NAC response should be based on biomarkers in combination rather than in isolation. Our study provides compelling evidence to support the use of ML-based models to predict NAC response in patients with TNBC.

Non-muscle-invasive bladder cancer molecular subtypes predict differential response to intravesical Bacillus Calmette-Guérin.

The recommended treatment for patients with high-risk non-muscle-invasive bladder cancer (HR-NMIBC) is tumor resection followed by adjuvant Bacillus Calmette-Guérin (BCG) bladder instillations. However, only 50% of patients benefit from this therapy. If progression to advanced disease occurs, then patients must undergo a radical cystectomy with risks of substantial morbidity and poor clinical outcome. Identifying tumors unlikely to respond to BCG can translate into alternative treatments, such as early radical cystectomy, targeted therapies, or immunotherapies. Here, we conducted molecular profiling of 132 patients with BCG-naive HR-NMIBC and 44 patients with recurrences after BCG (34 matched), which uncovered three distinct BCG response subtypes (BRS1, 2 and BRS3). Patients with BRS3 tumors had a reduced recurrence-free and progression-free survival compared with BRS1/2. BRS3 tumors expressed high epithelial-to-mesenchymal transition and basal markers and had an immunosuppressive profile, which was confirmed with spatial proteomics. Tumors that recurred after BCG were enriched for BRS3. BRS stratification was validated in a second cohort of 151 BCG-naive patients with HR-NMIBC, and the molecular subtypes outperformed guideline-recommended risk stratification based on clinicopathological variables. For clinical application, we confirmed that a commercially approved assay was able to predict BRS3 tumors with an area under the curve of 0.87. These BCG response subtypes will allow for improved identification of patients with HR-NMIBC at the highest risk of progression and have the potential to be used to select more appropriate treatments for patients unlikely to respond to BCG.

Impact of KRAS and BRAF mutations on treatment efficacy and survival in high-grade gastroenteropancreatic neuroendocrine neoplasms.

High-grade gastroenteropancreatic neuroendocrine neoplasms (HG GEP-NEN) typically disseminate early. Treatment of metastatic disease has limited benefit and prognosis is generally discouraging. Data on the clinical impact of mutations in HG GEP-NEN are scarce. There is an unmet need for reliable biomarkers to predict treatment outcome and prognosis in metastatic HG GEP-NEN. Patients with metastatic HG GEP-NEN diagnosed at three centres were selected for KRAS-, BRAF mutation and microsatellite instability (MSI) analyses. Results were linked to treatment outcome and overall survival. After pathological re-evaluation, 83 patients met inclusion criteria: 77 (93%) GEP neuroendocrine carcinomas (NEC) and six (7%) GEP neuroendocrine tumours (NET) G3. NEC harboured higher frequency of mutations than NET G3. Colon NEC harboured a particular high frequency of BRAF mutations (63%). Immediate disease progression on first-line chemotherapy was significantly higher for NEC with BRAF mutation (73%) versus wild-type (27%) (p = .016) and for colonic primary (65%) versus other NEC (28%) (p = .011). Colon NEC had a significant shorter PFS compared to other primary sites, a finding independent of BRAF status. Immediate disease progression was particularly frequent for BRAF mutated colon NEC (OR 10.2, p = .007). Surprisingly, BRAF mutation did not influence overall survival. KRAS mutation was associated with inferior overall survival for the whole NEC population (HR 2.02, p = .015), but not for those given first-line chemotherapy. All long-term survivors (>24 m) were double wild-type. Three NEC cases (4.8%) were MSI. Colon NEC with BRAF mutation predicted immediate disease progression on first-line chemotherapy, but did not affect PFS or OS. Benefit of first-line platinum/etoposide treatment seems limited for colon NEC, especially for BRAF mutated cases. KRAS mutations did not influence treatment efficacy nor survival for patients receiving first-line chemotherapy. Both frequency and clinical impact of KRAS/BRAF mutations in digestive NEC differ from prior results on digestive adenocarcinoma.

Automated Prognostic Assessment of Endometrial Hyperplasia for Progression Risk Evaluation Using Artificial Intelligence.

Endometrial hyperplasia is a precursor to endometrial cancer, characterized by excessive proliferation of glands that is distinguishable from normal endometrium. Current classifications define 2 types of EH, each with a different risk of progression to endometrial cancer. However, these schemes are based on visual assessments and, therefore, subjective, possibly leading to overtreatment or undertreatment. In this study, we developed an automated artificial intelligence tool (ENDOAPP) for the measurement of morphologic and cytologic features of endometrial tissue using the software Visiopharm. The ENDOAPP was used to extract features from whole-slide images of PAN-CK+-stained formalin-fixed paraffin-embedded tissue sections from 388 patients diagnosed with endometrial hyperplasia between 1980 and 2007. Follow-up data were available for all patients (mean = 140 months). The most prognostic features were identified by a logistic regression model and used to assign a low-risk or high-risk progression score. Performance of the ENDOAPP was assessed for the following variables: images from 2 different scanners (Hamamatsu XR and S60) and automated placement of a region of interest versus manual placement by an operator. Then, the performance of the application was compared with that of current classification schemes: WHO94, WHO20, and EIN, and the computerized-morphometric risk classification method: D-score. The most significant prognosticators were percentage stroma and the standard deviation of the lesser diameter of epithelial nuclei. The ENDOAPP had an acceptable discriminative power with an area under the curve of 0.765. Furthermore, strong to moderate agreement was observed between manual operators (intraclass correlation coefficient: 0.828) and scanners (intraclass correlation coefficient: 0.791). Comparison of the prognostic capability of each classification scheme revealed that the ENDOAPP had the highest accuracy of 88%-91% alongside the D-score method (91%). The other classification schemes had an accuracy between 83% and 87%. This study demonstrated the use of computer-aided prognosis to classify progression risk in EH for improved patient treatment.

Predicting Neoadjuvant Treatment Response in Triple-Negative Breast Cancer Using Machine Learning.

BACKGROUND: Neoadjuvant chemotherapy (NAC) is the standard treatment for early-stage triple negative breast cancer (TNBC). The primary endpoint of NAC is a pathological complete response (pCR). NAC results in pCR in only 30-40% of TNBC patients. Tumor-infiltrating lymphocytes (TILs), Ki67 and phosphohistone H3 (pH3) are a few known biomarkers to predict NAC response. Currently, systematic evaluation of the combined value of these biomarkers in predicting NAC response is lacking. In this study, the predictive value of markers derived from H&E and IHC stained biopsy tissue was comprehensively evaluated using a supervised machine learning (ML)-based approach. Identifying predictive biomarkers could help guide therapeutic decisions by enabling precise stratification of TNBC patients into responders and partial or non-responders. METHODS: Serial sections from core needle biopsies (n = 76) were stained with H&E and immunohistochemically for the Ki67 and pH3 markers, followed by whole-slide image (WSI) generation. The serial section stains in H&E stain, Ki67 and pH3 markers formed WSI triplets for each patient. The resulting WSI triplets were co-registered with H&E WSIs serving as the reference. Separate mask region-based CNN (MRCNN) models were trained with annotated H&E, Ki67 and pH3 images for detecting tumor cells, stromal and intratumoral TILs (sTILs and tTILs), Ki67+, and pH3+ cells. Top image patches with a high density of cells of interest were identified as hotspots. Best classifiers for NAC response prediction were identified by training multiple ML models and evaluating their performance by accuracy, area under curve, and confusion matrix analyses. RESULTS: Highest prediction accuracy was achieved when hotspot regions were identified by tTIL counts and each hotspot was represented by measures of tTILs, sTILs, tumor cells, Ki67+, and pH3+ features. Regardless of the hotspot selection metric, a complementary use of multiple histological features (tTILs, sTILs) and molecular biomarkers (Ki67 and pH3) resulted in top ranked performance at the patient level. CONCLUSIONS: Overall, our results emphasize that prediction models for NAC response should be based on biomarkers in combination rather than in isolation. Our study provides compelling evidence to support the use of ML-based models to predict NAC response in patients with TNBC.

Estimating diagnostic uncertainty in artificial intelligence assisted pathology using conformal prediction.

Unreliable predictions can occur when an artificial intelligence (AI) system is presented with data it has not been exposed to during training. We demonstrate the use of conformal prediction to detect unreliable predictions, using histopathological diagnosis and grading of prostate biopsies as example. We digitized 7788 prostate biopsies from 1192 men in the STHLM3 diagnostic study, used for training, and 3059 biopsies from 676 men used for testing. With conformal prediction, 1 in 794 (0.1%) predictions is incorrect for cancer diagnosis (compared to 14 errors [2%] without conformal prediction) while 175 (22%) of the predictions are flagged as unreliable when the AI-system is presented with new data from the same lab and scanner that it was trained on. Conformal prediction could with small samples (N = 49 for external scanner, N = 10 for external lab and scanner, and N = 12 for external lab, scanner and pathology assessment) detect systematic differences in external data leading to worse predictive performance. The AI-system with conformal prediction commits 3 (2%) errors for cancer detection in cases of atypical prostate tissue compared to 44 (25%) without conformal prediction, while the system flags 143 (80%) unreliable predictions. We conclude that conformal prediction can increase patient safety of AI-systems.

A Case Series Exploration of Multi-Regional Expression Heterogeneity in Triple-Negative Breast Cancer Patients.

Extensive intratumoral heterogeneity (ITH) is believed to contribute to therapeutic failure and tumor recurrence, as treatment-resistant cell clones can survive and expand. However, little is known about ITH in triple-negative breast cancer (TNBC) because of the limited number of single-cell sequencing studies on TNBC. In this study, we explored ITH in TNBC by evaluating gene expression-derived and imaging-derived multi-region differences within the same tumor. We obtained tissue specimens from 10 TNBC patients and conducted RNA sequencing analysis of 2-4 regions per tumor. We developed a novel analysis framework to dissect and characterize different types of variability: between-patients (inter-tumoral heterogeneity), between-patients across regions (inter-tumoral and region heterogeneity), and within-patient, between-regions (regional intratumoral heterogeneity). We performed a Bayesian changepoint analysis to assess and classify regional variability as low (convergent) versus high (divergent) within each patient feature (TNBC and PAM50 subtypes, immune, stroma, tumor counts and tumor infiltrating lymphocytes). Gene expression signatures were categorized into three types of variability: between-patients (108 genes), between-patients across regions (183 genes), and within-patients, between-regions (778 genes). Based on the between-patient gene signature, we identified two distinct patient clusters that differed in menopausal status. Significant intratumoral divergence was observed for PAM50 classification, tumor cell counts, and tumor-infiltrating T cell abundance. Other features examined showed a representation of both divergent and convergent results. Lymph node stage was significantly associated with divergent tumors. Our results show extensive intertumoral heterogeneity and regional ITH in gene expression and image-derived features in TNBC. Our findings also raise concerns regarding gene expression based TNBC subtyping. Future studies are warranted to elucidate the role of regional heterogeneity in TNBC as a driver of treatment resistance.

A spatial attention guided deep learning system for prediction of pathological complete response using breast cancer histopathology images.

MOTIVATION: Predicting pathological complete response (pCR) to neoadjuvant chemotherapy (NAC) in triple-negative breast cancer (TNBC) patients accurately is direly needed for clinical decision making. pCR is also regarded as a strong predictor of overall survival. In this work, we propose a deep learning system to predict pCR to NAC based on serial pathology images stained with hematoxylin and eosin and two immunohistochemical biomarkers (Ki67 and PHH3). To support human prior domain knowledge-based guidance and enhance interpretability of the deep learning system, we introduce a human knowledge-derived spatial attention mechanism to inform deep learning models of informative tissue areas of interest. For each patient, three serial breast tumor tissue sections from biopsy blocks were sectioned, stained in three different stains and integrated. The resulting comprehensive attention information from the image triplets is used to guide our prediction system for prognostic tissue regions. RESULTS: The experimental dataset consists of 26 419 pathology image patches of 1000×1000 pixels from 73 TNBC patients treated with NAC. Image patches from randomly selected 43 patients are used as a training dataset and images patches from the rest 30 are used as a testing dataset. By the maximum voting from patch-level results, our proposed model achieves a 93% patient-level accuracy, outperforming baselines and other state-of-the-art systems, suggesting its high potential for clinical decision making. AVAILABILITY AND IMPLEMENTATION: The codes, the documentation and example data are available on an open source at: https://github.com/jkonglab/PCR_Prediction_Serial_WSIs_biomarkers. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Liquid biopsies and patient-reported outcome measures for integrative monitoring of patients with early-stage breast cancer: a study protocol for the longitudinal observational Prospective Breast Cancer Biobanking (PBCB) study.

INTRODUCTION: Breast cancer is still the most common malignancy among women worldwide. The Prospective Breast Cancer Biobank (PBCB) collects blood and urine from patients with breast cancer every 6 or 12 months for 11 years from 2011 to 2030 at two university hospitals in Western Norway. The project aims to identify new biomarkers that enable detection of systemic recurrences at the molecular level. As blood represents the biological interface between the primary tumour, the microenvironment and distant metastases, liquid biopsies represent the ideal medium to monitor the patient's cancer biology for identification of patients at high risk of relapse and for early detection systemic relapse.Including patient-reported outcome measures (PROMs) allows for a vast number of possibilities to compare PROM data with biological information, enabling the study of fatigue and Quality of Life in patients with breast cancer. METHODS AND ANALYSIS: A total of 1455 patients with early-stage breast cancer are enrolled in the PBCB study, which has a one-armed prospective observational design. Participants consent to contribute liquid biopsies (i.e., peripheral blood and urine samples) every 6 or 12 months for 11 years. The liquid biopsies are the basis for detection of circulating tumour cells, circulating tumour DNA (ctDNA), exosomal micro-RNA (miRNA), miRNA in Tumour Educated Platelet and metabolomic profiles. In addition, participants respond to 10 PROM questionnaires collected annually. Moreover, a control group comprising 200 women without cancer aged 25-70 years will provide the same data. ETHICS AND DISSEMINATION: The general research biobank PBCB was approved by the Ministry of Health and Care Services in 2007, by the Regional Ethics Committee (REK) in 2010 (#2010/1957). The PROM (#2011/2161) and the biomarker study PerMoBreCan (#2015/2010) were approved by REK in 2011 and 2015 respectively. Results will be published in international peer reviewed journals. Deidentified data will be accessible on request. TRIAL REGISTRATION NUMBER: NCT04488614.

A Gene Signature Identifying CIN3 Regression and Cervical Cancer Survival.

The purpose of this study was to establish a gene signature that may predict CIN3 regression and that may aid in selecting patients who may safely refrain from conization. Oncomine mRNA data including 398 immune-related genes from 21 lesions with confirmed regression and 28 with persistent CIN3 were compared. L1000 mRNA data from a cervical cancer cohort was available for validation (n = 239). Transcriptomic analyses identified TDO2 (p = 0.004), CCL5 (p < 0.001), CCL3 (p = 0.04), CD38 (p = 0.02), and PRF1 (p = 0.005) as upregulated, and LCK downregulated (p = 0.01) in CIN3 regression as compared to persistent CIN3 lesions. From these, a gene signature predicting CIN3 regression with a sensitivity of 91% (AUC = 0.85) was established. Transcriptomic analyses revealed proliferation as significantly linked to persistent CIN3. Within the cancer cohort, high regression signature score associated with immune activation by Gene Set enrichment Analyses (GSEA) and immune cell infiltration by histopathological evaluation (p < 0.001). Low signature score was associated with poor survival (p = 0.007) and large tumors (p = 0.01). In conclusion, the proposed six-gene signature predicts CIN regression and favorable cervical cancer prognosis and points to common drivers in precursors and cervical cancer lesions.

Crown-Like Structures in Breast Adipose Tissue: Early Evidence and Current Issues in Breast Cancer.

Obesity is an established risk factor for postmenopausal breast cancer and has been linked to worse breast cancer prognosis, most clearly for hormone receptor-positive breast cancers. The underlying mechanisms of the obesity-breast cancer association are not fully understood, but growing evidence points to the breast adipose tissue microenvironment playing an important role. Obesity-induced adipose tissue dysfunction can result in a chronic state of low-grade inflammation. Crown-like structures of the breast (CLS-B) were recently identified as a histologic marker of local inflammation. In this review, we evaluate the early evidence of CLS-B in breast cancer. Data from preclinical and clinical studies show that these inflammatory lesions within the breast are associated with local NF-κB activation, increased aromatase activity, and elevation of pro-inflammatory mediators (TNFα, IL-1ß, IL-6, and COX-2-derived PGE2)-factors involved in multiple pathways of breast cancer development and progression. There is also substantial evidence from epidemiologic studies that CLS-B are associated with greater adiposity among breast cancer patients. However, there is insufficient evidence that CLS-B impact breast cancer risk or prognosis. Comparisons across studies of prognosis were complicated by differences in CLS-B evaluation and deficiencies in study design, which future studies should take into consideration. Breast adipose tissue inflammation provides a plausible explanation for the obesity-breast cancer association, but further study is needed to establish its role and whether markers such as CLS-B are clinically useful.

miRNA normalization enables joint analysis of several datasets to increase sensitivity and to reveal novel miRNAs differentially expressed in breast cancer.

Different miRNA profiling protocols and technologies introduce differences in the resulting quantitative expression profiles. These include differences in the presence (and measurability) of certain miRNAs. We present and examine a method based on quantile normalization, Adjusted Quantile Normalization (AQuN), to combine miRNA expression data from multiple studies in breast cancer into a single joint dataset for integrative analysis. By pooling multiple datasets, we obtain increased statistical power, surfacing patterns that do not emerge as statistically significant when separately analyzing these datasets. To merge several datasets, as we do here, one needs to overcome both technical and batch differences between these datasets. We compare several approaches for merging and jointly analyzing miRNA datasets. We investigate the statistical confidence for known results and highlight potential new findings that resulted from the joint analysis using AQuN. In particular, we detect several miRNAs to be differentially expressed in estrogen receptor (ER) positive versus ER negative samples. In addition, we identify new potential biomarkers and therapeutic targets for both clinical groups. As a specific example, using the AQuN-derived dataset we detect hsa-miR-193b-5p to have a statistically significant over-expression in the ER positive group, a phenomenon that was not previously reported. Furthermore, as demonstrated by functional assays in breast cancer cell lines, overexpression of hsa-miR-193b-5p in breast cancer cell lines resulted in decreased cell viability in addition to inducing apoptosis. Together, these observations suggest a novel functional role for this miRNA in breast cancer. Packages implementing AQuN are provided for Python and Matlab: https://github.com/YakhiniGroup/PyAQN.

High-Grade Cervical Intraepithelial Neoplasia (CIN) Associates with Increased Proliferation and Attenuated Immune Signaling.

Implementation of high-risk human papilloma virus (HPV) screening and the increasing proportion of HPV vaccinated women in the screening program will reduce the percentage of HPV positive women with oncogenic potential. In search of more specific markers to identify women with high risk of cancer development, we used RNA sequencing to compare the transcriptomic immune-profile of 13 lesions with cervical intraepithelial neoplasia grade 3 (CIN3) or adenocarcinoma in situ (AIS) and 14 normal biopsies from women with detected HPV infections. In CIN3/AIS lesions as compared to normal tissue, 27 differential expressed genes were identified. Transcriptomic analysis revealed significantly higher expression of a number of genes related to proliferation, (CDKN2A, MELK, CDK1, MKI67, CCNB2, BUB1, FOXM1, CDKN3), but significantly lower expression of genes related to a favorable immune response (NCAM1, ARG1, CD160, IL18, CX3CL1). Compared to the RNA sequencing results, good correlation was achieved with relative quantitative PCR analysis for NCAM1 and CDKN2A. Quantification of NCAM1 positive cells with immunohistochemistry showed epithelial reduction of NCAM1 in CIN3/AIS lesions. In conclusion, NCAM1 and CDKN2A are two promising candidates to distinguish whether women are at high risk of developing cervical cancer and in need of frequent follow-up.