Outcome and Biomarker Supervised Deep Learning for Survival Prediction in Two Multicenter Breast Cancer Series.

BACKGROUND: Prediction of clinical outcomes for individual cancer patients is an important step in the disease diagnosis and subsequently guides the treatment and patient counseling. In this work, we develop and evaluate a joint outcome and biomarker supervised (estrogen receptor expression and ERBB2 expression and gene amplification) multitask deep learning model for prediction of outcome in breast cancer patients in two nation-wide multicenter studies in Finland (the FinProg and FinHer studies). Our approach combines deep learning with expert knowledge to provide more accurate, robust, and integrated prediction of breast cancer outcomes. MATERIALS AND METHODS: Using deep learning, we trained convolutional neural networks (CNNs) with digitized tissue microarray (TMA) samples of primary hematoxylin-eosin-stained breast cancer specimens from 693 patients in the FinProg series as input and breast cancer-specific survival as the endpoint. The trained algorithms were tested on 354 TMA patient samples in the same series. An independent set of whole-slide (WS) tumor samples from 674 patients in another multicenter study (FinHer) was used to validate and verify the generalization of the outcome prediction based on CNN models by Cox survival regression and concordance index (c-index). Visual cancer tissue characterization, i.e., number of mitoses, tubules, nuclear pleomorphism, tumor-infiltrating lymphocytes, and necrosis was performed on TMA samples in the FinProg test set by a pathologist and combined with deep learning-based outcome prediction in a multitask algorithm. RESULTS: The multitask algorithm achieved a hazard ratio (HR) of 2.0 (95% confidence interval [CI] 1.30-3.00), P < 0.001, c-index of 0.59 on the 354 test set of FinProg patients, and an HR of 1.7 (95% CI 1.2-2.6), P = 0.003, c-index 0.57 on the WS tumor samples from 674 patients in the independent FinHer series. The multitask CNN remained a statistically independent predictor of survival in both test sets when adjusted for histological grade, tumor size, and axillary lymph node status in a multivariate Cox analyses. An improved accuracy (c-index 0.66) was achieved when deep learning was combined with the tissue characteristics assessed visually by a pathologist. CONCLUSIONS: A multitask deep learning algorithm supervised by both patient outcome and biomarker status learned features in basic tissue morphology predictive of survival in a nationwide, multicenter series of patients with breast cancer. The algorithms generalized to another independent multicenter patient series and whole-slide breast cancer samples and provide prognostic information complementary to that of a comprehensive series of established prognostic factors.

Deep learning identifies morphological features in breast cancer predictive of cancer ERBB2 status and trastuzumab treatment efficacy.

The treatment of patients with ERBB2 (HER2)-positive breast cancer with anti-ERBB2 therapy is based on the detection of ERBB2 gene amplification or protein overexpression. Machine learning (ML) algorithms can predict the amplification of ERBB2 based on tumor morphological features, but it is not known whether ML-derived features can predict survival and efficacy of anti-ERBB2 treatment. In this study, we trained a deep learning model with digital images of hematoxylin-eosin (H&E)-stained formalin-fixed primary breast tumor tissue sections, weakly supervised by ERBB2 gene amplification status. The gene amplification was determined by chromogenic in situ hybridization (CISH). The training data comprised digitized tissue microarray (TMA) samples from 1,047 patients. The correlation between the deep learning-predicted ERBB2 status, which we call H&E-ERBB2 score, and distant disease-free survival (DDFS) was investigated on a fully independent test set, which included whole-slide tumor images from 712 patients with trastuzumab treatment status available. The area under the receiver operating characteristic curve (AUC) in predicting gene amplification in the test sets was 0.70 (95% CI, 0.63-0.77) on 354 TMA samples and 0.67 (95% CI, 0.62-0.71) on 712 whole-slide images. Among patients with ERBB2-positive cancer treated with trastuzumab, those with a higher than the median morphology-based H&E-ERBB2 score derived from machine learning had more favorable DDFS than those with a lower score (hazard ratio [HR] 0.37; 95% CI, 0.15-0.93; P = 0.034). A high H&E-ERBB2 score was associated with unfavorable survival in patients with ERBB2-negative cancer as determined by CISH. ERBB2-associated morphology correlated with the efficacy of adjuvant anti-ERBB2 treatment and can contribute to treatment-predictive information in breast cancer.

Breast cancer outcome prediction with tumour tissue images and machine learning.

PURPOSE: Recent advances in machine learning have enabled better understanding of large and complex visual data. Here, we aim to investigate patient outcome prediction with a machine learning method using only an image of tumour sample as an input. METHODS: Utilising tissue microarray (TMA) samples obtained from the primary tumour of patients (N = 1299) within a nationwide breast cancer series with long-term-follow-up, we train and validate a machine learning method for patient outcome prediction. The prediction is performed by classifying samples into low or high digital risk score (DRS) groups. The outcome classifier is trained using sample images of 868 patients and evaluated and compared with human expert classification in a test set of 431 patients. RESULTS: In univariate survival analysis, the DRS classification resulted in a hazard ratio of 2.10 (95% CI 1.33-3.32, p = 0.001) for breast cancer-specific survival. The DRS classification remained as an independent predictor of breast cancer-specific survival in a multivariate Cox model with a hazard ratio of 2.04 (95% CI 1.20-3.44, p = 0.007). The accuracy (C-index) of the DRS grouping was 0.60 (95% CI 0.55-0.65), as compared to 0.58 (95% CI 0.53-0.63) for human expert predictions based on the same TMA samples. CONCLUSIONS: Our findings demonstrate the feasibility of learning prognostic signals in tumour tissue images without domain knowledge. Although further validation is needed, our study suggests that machine learning algorithms can extract prognostically relevant information from tumour histology complementing the currently used prognostic factors in breast cancer.

The presence of sinusoidal CD163(+) macrophages in lymph nodes is associated with favorable nodal status in patients with breast cancer.

As macrophages are some of the first cells to encounter metastatic tumor cells in sentinel lymph nodes (SLN) and natural killer (NK) cells are critical to the cytotoxicity of abnormal cells, we sought to determine if these cell populations were altered in the presence of nodal metastasis. We used immunohistochemistry to assess the SLN of 47 patients with breast cancer (36 with nodal metastasis and 11 without nodal metastasis) and 10 control lymph nodes. We assessed metastatic areas and nonmetastatic areas separately for CD163, a marker of macrophages, and ANK-1, a marker for precursors of activated NK cells. Positively stained cells were manually counted in multiple high-power fields and averaged. Groups were compared with the Kruskal-Wallis test. Spearman rank order test was used for correlations. There was a lower frequency of CD163(+) macrophages in the SLN of patients with breast cancer (median, 11.0 %; range, 4.1-20.4 %) than controls (median, 16.5 %; range, 8.9-19.6 %; p = 0.002). There were no differences in the expression of ANK between patients with cancer (median, 1.4 %; range, 0.23-6.3 %) and controls (median, 1.5 %; range, 0.60-5.4 %; p = 0.5). In patients with nodal metastasis, the accumulation of CD163(+) cells in the sinuses correlated negatively with CD8(+) tumor-infiltrating lymphocytes (r (2) = 0.23; p = 0.001). These results suggest that the reduction of CD163(+) macrophages in the sinuses of the SLN is associated with nodal metastasis and may have a role in regional immunity.

Metastasis to sentinel lymph nodes in breast cancer is associated with maturation arrest of dendritic cells and poor co-localization of dendritic cells and CD8+ T cells.

The regional immune systems of patients with breast cancer are immunosuppressed. Dendritic cells are professional antigen-presenting cells and present cancer-associated antigens to the adaptive immune system in sentinel lymph nodes. Dendritic cells may promote, or inhibit, an adaptive immune response to specific antigens. Our aim was to assess whether dendritic cells were associated with nodal metastasis in patients with breast cancer. Sentinel lymph nodes of 47 patients with breast cancer with varying degrees of nodal disease and ten controls were evaluated using immunohistochemistry for the accumulation of dendritic cells in general (CD1a(+)), mature dendritic cells (CD208(+)), and plasmacytoid dendritic cells (CD123(+)). Cytotoxic T cell and regulatory T cell accumulation were also evaluated. Sentinel lymph nodes with macrometastases demonstrated fewer mature dendritic cells than sentinel lymph nodes without metastasis (p = 0.028), but not controls. There were fewer mature dendritic cells to cytotoxic T cells in sentinel lymph nodes with metastasis than those without (p = 0.033). Also, there were more regulatory T cells to mature dendritic cells in sentinel lymph nodes with metastasis than those without (p = 0.02). In conclusion, our study suggests that sentinel lymph nodes with metastasis have arrest of maturation of dendritic cells, fewer mature dendritic cell interactions with cytotoxic T cells, and more regulatory T cells than sentinel lymph nodes without metastasis in patients with breast cancer. These findings extend our understanding of regional immunosuppression and suggest that most regional immunosuppressive changes are associated with nodal metastasis in breast cancer.

Comparison of in vitro breast cancer visibility in analyser-based computed tomography with histopathology, mammography, computed tomography and magnetic resonance imaging.

High-resolution analyser-based X-ray imaging computed tomography (HR ABI-CT) findings on in vitro human breast cancer are compared with histopathology, mammography, computed tomography (CT) and magnetic resonance imaging. The HR ABI-CT images provided significantly better low-contrast visibility compared with the standard radiological images. Fine cancer structures indistinguishable and superimposed in mammograms were seen, and could be matched with the histopathological results. The mean glandular dose was less than 1 mGy in mammography and 12-13 mGy in CT and ABI-CT. The excellent visibility of in vitro breast cancer suggests that HR ABI-CT may have a valuable role in the future as an adjunct or even alternative to current breast diagnostics, when radiation dose is further decreased, and compact synchrotron radiation sources become available.

Metastatic outgrowth encompasses COL-I, FN1, and POSTN up-regulation and assembly to fibrillar networks regulating cell adhesion, migration, and growth.

Although the outgrowth of micrometastases into macrometastases is the rate-limiting step in metastatic progression and the main determinant of cancer fatality, the molecular mechanisms involved have been little studied. Here, we compared the gene expression profiles of melanoma lymph node micro- and macrometastases and unexpectedly found no common up-regulation of any single growth factor/cytokine, except for the cytokine-like SPP1. Importantly, metastatic outgrowth was found to be consistently associated with activation of the transforming growth factor-beta signaling pathway (confirmed by phospho-SMAD2 staining) and concerted up-regulation of POSTN, FN1, COL-I, and VCAN genes-all inducible by transforming growth factor-beta. The encoded extracellular matrix proteins were found to together form intricate fibrillar networks around tumor cell nests in melanoma and breast cancer metastases from various organs. Functional analyses suggested that these newly synthesized protein networks regulate adhesion, migration, and growth of tumor cells, fibroblasts, and endothelial cells. POSTN acted as an anti-adhesive molecule counteracting the adhesive functions of FN1 and COL-I. Further, cellular FN and POSTN were specifically overexpressed in the newly forming/formed tumor blood vessels. Transforming growth factor-beta receptors and the metastasis-related matrix proteins, POSTN and FN1, in particular, may thus provide attractive targets for development of new therapies against disseminated melanoma, breast cancer, and possibly other tumors, by affecting key processes of metastasis: tumor/stromal cell migration, growth, and angiogenesis.

Concordance between HER-2 and steroid hormone receptor expression between primary breast cancer, sentinel node metastases, and isolated tumor cells.

Our aim was to compare the HER-2 and hormone receptor status between primary tumor (PT) and sentinel node (SN) metastases in breast cancer. We analyzed 65 PTs with 42 macrometastases, 18 micrometastases, and 5 ITCs for the HER-2 amplification status and for the hormone receptor status. In HER-2-positive PTs, 26 of 29 metastases were HER-2-positive, including all six analyzed micrometastases and ITCs. Three macrometastases were HER-2-negative. In HER-2-negative PTs, 35 of 36 metastases were HER-2-negative. A low-grade HER-2 amplification was detected in one micrometastasis. A full concordance between the PT and SN metastases in both HER-2 amplification and ER and PR status was observed in 40 of 56 (71%) cases. These 40 fully concordant cases included 27 macrometastatic, 9 micrometastatic, and 4 ITC-cases. The concordance in HER-2, ER, and PR was high between the PT and their SN metastases. However, the concordance may be remarkably lower when analyzed simultaneously for all three predictive factors.

Simultaneous Foxp3 and IDO expression is associated with sentinel lymph node metastases in breast cancer.

BACKGROUND: There is evidence that the immune systems of patients with breast cancer are dysfunctional. Regulatory T cells (Tregs), and IDO, an immunosuppressive enzyme, are associated with more advanced disease in some cancers and may promote immunologic tolerance to tumors. Our aim was to assess whether expression of Foxp3, a marker of Tregs, and IDO were linked with nodal metastasis in breast cancer patients. Inhibitors of IDO are available and could potentially demonstrate utility in breast cancer if IDO drives progression of disease. METHODS: Sentinel lymph nodes (SLN) of 47 breast cancer patients with varying degrees of nodal disease and 10 controls were evaluated for expression of Foxp3 and IDO using immunohistochemistry. Positively stained cells were quantified and their distribution within the SLN noted. RESULTS: The proportion of Foxp3+ cells was higher in SLN of cancer patients than controls (19% v. 10%, p < 0.001). Specifically, there were more Foxp3+ cells in SLN with metastasis than tumor-free SLN (20% v. 14%, p = 0.02). The proportion IDO+ cell in SLN of cancer patients was not statistically different than controls (4.0% v. 1.6%, p = 0.08). In order to demonstrate the combined immunosuppressive effect of Foxp3 and IDO, we categorized each SLN as positive or negative for Foxp3 and IDO. The Foxp3+/IDO+ group almost exclusively consisted of cancer patients with node positive disease. CONCLUSION: In conclusion, our study shows that Foxp3+ cells are associated with more advanced disease in breast cancer, a finding that is proving to be true in many other cancers. As IDO has been found to promote differentiation of Tregs, IDO may become a suitable target to abrogate the development of T-cell tolerance and to promote an effective immune response to breast cancer. Our results about the combined expression of IDO and Foxp3 in metastastic SLN support this assumption.

The outcome of sentinel node biopsy in breast cancer patients with preoperative surgical biopsy.

BACKGROUND: The aim of the study was to evaluate the outcome of sentinel node biopsy (SNB), especially the medium term axillary recurrence rate after negative SNB in patients with preoperative surgical biopsy (SB). PATIENTS AND METHODS: The study included 1,641 patients with a histological stage T1 tumours and SNB. In 77 patients, the diagnosis was obtained with SB, while 1,564 patients had underwent needle biopsy (NB) only. Axillary clearance was omitted in 56 SB patients and 921 NB patients after negative SNB. The median duration of follow-up in these patients was 54 months. RESULTS: None of the SB patients had axillary recurrences during the follow-up. Six NB patients had isolated axillary recurrences while three patients had concomitant local and axillary recurrences. There were no differences in local or distant recurrences or breast cancer deaths between the SB and the NB patients. CONCLUSIONS: SNB seems a feasible axillary staging method in patients with histological stage T1 tumour also after preoperative SB. The risk of axillary recurrence after negative SNB is negligible in these patients.

Germ-line variation at a functional p53 binding site increases susceptibility to breast cancer development.

UNLABELLED: Multiple lines of evidence suggest regulatory variation to play an important role in phenotypic evolution and disease development, but few regulatory polymorphisms have been characterized genetically and molecularly. Recent technological advances have made it possible to identify bona fide regulatory sequences experimentally on a genome-wide scale and opened the window for the biological interrogation of germ-line polymorphisms within these sequences. In this study, through a forward genetic analysis of bona fide p53 binding sites identified by a genome-wide chromatin immunoprecipitation and sequence analysis, we discovered a SNP (rs1860746) within the motif sequence of a p53 binding site where p53 can function as a regulator of transcription. We found that the minor allele (T) binds p53 poorly and has low transcriptional regulation activity as compared to the major allele (G). Significantly, the homozygosity of the minor allele was found to be associated with an increased risk of ER negative breast cancer (OR = 1.47, P = 0.038) from the analysis of five independent breast cancer samples of European origin consisting of 6,127 breast cancer patients and 5,197 controls. rs1860746 resides in the third intron of the PRKAG2 gene that encodes the γ subunit of the AMPK protein, a major sensor of metabolic stress and a modulator of p53 action. However, this gene does not appear to be regulated by p53 in lymphoblastoid cell lines nor in a cancer cell line. These results suggest that either the rs1860746 locus regulates another gene through distant interactions, or that this locus is in linkage disequilibrium with a second causal mutation. This study shows the feasibility of using genomic scale molecular data to uncover disease associated SNPs, but underscores the complexity of determining the function of regulatory variants in human populations. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (doi:10.1007/s11568-010-9138-x) contains supplementary material, which is available to authorized users.

Margin status after breast-conserving treatment of breast cancer: how much free margin is enough?

In breast conserving surgery the margins of the specimen have to be tumor free. There is no universal agreement on the width of the tumor free margin. The width has some impact on recurrence rate, and at least 5 mm seems to be preferred and this is especially important in extensive intraductal component. In ductal cancer in situ the free margin should be at least 5 mm, preferably 10 mm. Oncoplastic surgery is beneficial for reaching adequate margins.

Toward high-contrast breast CT at low radiation dose.

This study was approved by the local research ethics committee, and patient informed consent was obtained. The purpose of this study was to demonstrate that high-spatial-resolution low-dose analyzer-based x-ray computed tomography (CT) can substantially improve the radiographic contrast of breast tissue in vitro when compared with that attained by using diagnostic mammography and CT. An excised human breast tumor was examined by using analyzer-based x-ray imaging with synchrotron radiation. The correspondence between analyzer-based x-ray images and diagnostic mammograms, CT images, and histopathologic findings was determined. Calcifications and fine details of soft tissue, which are at the contrast detection limit on diagnostic mammograms, are clearly visible on planar analyzer-based x-ray images. Analyzer-based x-ray CT yields high contrast from smoothly varying internal structures, such as tumorous mass lesions, corresponding to information on actual structures seen at histopathologic analysis. The mean glandular dose of 1.9 mGy in analyzer-based x-ray CT is approximately equivalent to the dose administered during single-view screening mammography. The improved visibility of mammographically indistinguishable lesions in vitro suggests that analyzer-based x-ray CT may be a valuable method in radiographic evaluation of the breast, thereby justifying further investigations.

Unsuccessful preoperative biopsies, fine needle aspiration cytology or core needle biopsy, lead to increased costs in the diagnostic workup in breast cancer.

Correct preoperative diagnosis of a breast lesion is essential for optimal treatment planning. Our aim was to compare feasibility of fine needle aspiration cytology (FNAC) and core needle biopsy (CNB) in diagnosis of breast lesions. The special aim was to evaluate the extra costs and delay in surgical treatment due to unsuccessful preoperative biopsies. Diagnostic work-ups in 572 patients with 580 breast lesions were retrospectively evaluated. FNAC was the first biopsy method for 339 lesions, CNB for 241 lesions. The postoperative diagnosis was malignant for 503 lesions. The preoperative rate of definitely malignant diagnosis was 67% (194/289) for FNAC and 96% (206/214) for CNB (p < 0.0001), and 95% and 99%, respectively (p = 0.0173), when also suspicious findings were included. In patients with FNAC, an additional needle biopsy was performed for 93 and a surgical biopsy for 62 lesions. In the CNB group, a subsequent CNB was performed for 2 and a surgical biopsy for 33. The frequent need for additional biopsies raised the total expenses of FNAC over those of CNB. Multiple biopsies may also delay cancer surgery. It is therefore recommended to use CNB as the initial needle biopsy method.

NAD(P)H:quinone oxidoreductase 1 NQO1*2 genotype (P187S) is a strong prognostic and predictive factor in breast cancer.

NQO1 guards against oxidative stress and carcinogenesis and stabilizes p53. We find that a homozygous common missense variant (NQO1(*)2, rs1800566(T), NM_000903.2:c.558C>T) that disables NQO1 strongly predicts poor survival among two independent series of women with breast cancer (P = 0.002, N = 1,005; P = 0.005, N = 1,162), an effect particularly evident after anthracycline-based adjuvant chemotherapy with epirubicin (P = 7.52 x 10(-6)) and in p53-aberrant tumors (P = 6.15 x 10(-5)). Survival after metastasis was reduced among NQO1(*)2 homozygotes, further implicating NQO1 deficiency in cancer progression and treatment resistance. Consistently, response to epirubicin was impaired in NQO1(*)2-homozygous breast carcinoma cells in vitro, reflecting both p53-linked and p53-independent roles of NQO1. We propose a model of defective anthracycline response in NQO1-deficient breast tumors, along with increased genomic instability promoted by elevated reactive oxygen species (ROS), and suggest that the NQO1 genotype is a prognostic and predictive marker for breast cancer.

Heterogeneity of breast cancer associations with five susceptibility loci by clinical and pathological characteristics.

A three-stage genome-wide association study recently identified single nucleotide polymorphisms (SNPs) in five loci (fibroblast growth receptor 2 (FGFR2), trinucleotide repeat containing 9 (TNRC9), mitogen-activated protein kinase 3 K1 (MAP3K1), 8q24, and lymphocyte-specific protein 1 (LSP1)) associated with breast cancer risk. We investigated whether the associations between these SNPs and breast cancer risk varied by clinically important tumor characteristics in up to 23,039 invasive breast cancer cases and 26,273 controls from 20 studies. We also evaluated their influence on overall survival in 13,527 cases from 13 studies. All participants were of European or Asian origin. rs2981582 in FGFR2 was more strongly related to ER-positive (per-allele OR (95%CI) = 1.31 (1.27-1.36)) than ER-negative (1.08 (1.03-1.14)) disease (P for heterogeneity = 10(-13)). This SNP was also more strongly related to PR-positive, low grade and node positive tumors (P = 10(-5), 10(-8), 0.013, respectively). The association for rs13281615 in 8q24 was stronger for ER-positive, PR-positive, and low grade tumors (P = 0.001, 0.011 and 10(-4), respectively). The differences in the associations between SNPs in FGFR2 and 8q24 and risk by ER and grade remained significant after permutation adjustment for multiple comparisons and after adjustment for other tumor characteristics. Three SNPs (rs2981582, rs3803662, and rs889312) showed weak but significant associations with ER-negative disease, the strongest association being for rs3803662 in TNRC9 (1.14 (1.09-1.21)). rs13281615 in 8q24 was associated with an improvement in survival after diagnosis (per-allele HR = 0.90 (0.83-0.97). The association was attenuated and non-significant after adjusting for known prognostic factors. Our findings show that common genetic variants influence the pathological subtype of breast cancer and provide further support for the hypothesis that ER-positive and ER-negative disease are biologically distinct. Understanding the etiologic heterogeneity of breast cancer may ultimately result in improvements in prevention, early detection, and treatment.

Complications of skin-sparing mastectomy followed by immediate breast reconstruction: a prospective randomized study comparing high-frequency radiosurgery with conventional diathermy.

Skin-sparing mastectomy (SSM) followed by immediate breast reconstruction delivers superior cosmetic and functional outcome. However, SSM is vulnerable to complications of the native skin envelope. This study aims to compare the effects of radiofrequency coagulation and conventional diathermy on complications of SSM. Sixty consecutive patients suitable for SSM were randomized into conventional diathermy and radiosurgery groups. These groups were compared and the risk factors for SSM flap complications were evaluated. The SSM flap complication rate was 23.4%. There was no difference between the study groups regarding the SSM flap complications. Increased SSM flap complication rate was associated with smoking and the type of skin incision used. This study shows that high-frequency radiosurgery is comparable to conventional diathermy in terms of complication rates of SSM. Furthermore, this study reports an association between the tennis-racquet-type incision and an increased SSM flap complication rate compared with the round periareolar type incision.

VEGF-D in association with VEGFR-3 promotes nodal metastasis in human invasive lobular breast cancer.

We assessed the expression of vascular endothelial growth factors (VEGF-C and VEGF-D) in breast cancer cells and the density of lymph vessels and VEGF receptor-3 (VEGFR-3)-positive vessels in and around the tumor in invasive lobular breast cancer. We found significant correlation between peritumoral lymph vessel density and presence of lymph node metastases (P=.001) and the number of metastatic lymph nodes (P<.001). A significant correlation was detected between tumor cell VEGF-D expression and lymph node status (P=.001) and density of lymphatic vessel endothelial receptor (LYVE)-1-positive vessels (P=.035). VEGFR-3+/VEGF-D+ and VEGFR-3+/VEGF-C+ tumors had a significantly higher number of metastatic lymph nodes than tumors with other staining patterns (P<.001). Tumors positive for neither VEGF-D nor VEGFR-3 had a lower density of LYVE-1+ vessels than tumors with other staining patterns (P=.033). Our results indicate that peritumoral lymph vessel density is associated with lymph node metastases in invasive lobular breast cancer and that invasive lobular cancer producing VEGF-D, surrounded by VEGFR-3+ vessels, has a significantly higher peritumoral lymph vessel density and a higher number of metastatic lymph nodes.

Predicting invasion in patients with DCIS in the preoperative percutaneous biopsy.

When ductal carcinoma in situ (DCIS) is suspected in mammography, core needle biopsy or vacuum assisted biopsy is recommended. However, invasion remains undetected with percutaneous biopsy techniques in 10-20% of the patients. Our aim was to evaluate the prevalence of and predictive factors for invasion in the surgical specimen in patients with DCIS in the preoperative biopsy. Sixty-seven consecutive participants of the Helsinki City Mammography Screening program with DCIS in the preoperative percutaneous biopsy were included. The palpability, the mammographical size and appearance and the visibility of the lesion in breast ultrasound were evaluate as factors predictive for invasion, as well as the histopathological features of DCIS in the preoperative biopsy. Twenty patients had invasion in the surgical specimen. The only predictive factor for invasion was the visibility of the lesion in ultrasound, but even this finding failed to reach statistical significance. Thirteen of the 26 patients with lesions visible in US had invasion in their surgical specimens, while only seven of the 41 patients without such a lesion had invasive or microinvasive cancer, Pc = 0.0686. In conclusion, the visibility of the lesion in US may predict detecting invasion in the surgical specimen in patients with DCIS in the preoperative biopsy.

High-resolution CT by diffraction-enhanced x-ray imaging: mapping of breast tissue samples and comparison with their histo-pathology.

The aim of this study was to introduce high-resolution computed tomography (CT) of breast tumours using the diffraction-enhanced x-ray imaging (DEI) technique and to compare results with radiological and histo-pathological examinations. X-ray CT images of tumour-bearing breast tissue samples were acquired by monochromatic synchrotron radiation (SR). Due to the narrow beam and a large sample-to-detector distance scattering is rejected in the absorption contrast images (SR-CT). Large contrast enhancement is achieved by the use of the DEI-CT method, where the effects of refraction and scatter rejection are analysed by crystal optics. Clinical mammograms and CT images were recorded as reference material for a radiological examination. Three malignant and benign samples were studied in detail. Their radiographs were compared with optical images of stained histological sections. The DEI-CT images map accurately the morphology of the samples, including collagen strands and micro-calcifications of dimensions less than 0.1 mm. Histo-pathological examination and reading of the radiographs were done independently, and the conclusions were in general agreement. High-resolution DEI-CT images show strong contrast and permit visualization of details invisible in clinical radiographs. The radiation dose may be reduced by an order of magnitude without compromising image quality, which would make possible clinical in vivo DEI-CT with future compact SR sources.