514 Background: In breast cancer, the majority of clinically node positive (cN1) patients are treated with primary systemic treatment (PST), with increasing rates of axillary pathologic complete response (pCR). Evidence supporting tailoring axillary treatment according to the response is growing. However, it is still debatable whether ALND is beneficial in cN1 patients with residual nodal disease and if any further axillary treatment is necessary in patients achieving pCR. In this study, we present the four-year oncologic outcome of cN1 patients undergoing either axillary radiation treatment or no axillary treatment according to the ‘Marking Axillary lymph nodes with Radioactive Iodine seeds’ (MARI)-protocol. Methods: Between 2014 and 2021, we prospectively enrolled patients with node positive breast cancer and 1 - 3 involved axillary lymph nodes on FDG-PET/CT pre-PST, who were treated according to the MARI-protocol. Patients with residual disease of the MARI node (ypN1) received axillary radiotherapy (RT), while patients with a pCR of the MARI node (ypN0) received no further axillary treatment. The primary endpoint was the axillary recurrence rate. Secondary endpoints include invasive disease free survival (iDFS) and overall survival (OS), calculated using the Kaplan-Meier method. Results: Of the 367 included patients, more than half (54%, n= 198) had hormonal receptor positive (HR+)/Human Epidermal growth factor 2 negative (HER2-) breast cancer; 59 patients (16%) had HR+/HER2+; 39 patients (11%) had HR-/HER2+ and 71 patients (19%) had triple negative (TN) subtype. Median age was 49 (IQR 41 – 56) years. 231 (63%) patients who had ypN1 received RT and 136 patients (37%) who achieved ypN0 received no further axillary treatment. Median follow up was 49 (IQR 32 – 70) months. Axillary recurrence rate was 3.5% ( n= 8) in patients with ypN1 and 0.7% ( n= 1) in patients with ypN0. Most axillary recurrences ( n= 6) occurred simultaneously with distant metastases in the ypN1 group. In patients with ypN1, the iDFS was 86% and the OS 93%; in ypN0 patients, the iDFS was 93% and the OS 98%. Conclusions: Response-guided axillary treatment in patients with limited nodal disease according to the MARI protocol is associated with favorable oncologic outcome without performing ALND’s. De-escalation can therefore be considered safe. [Table: see text]

Background Neoadjuvant systemic treatment (NST) leads to pathologic complete response (pCR) in 10–89% of breast cancer patients depending on subtype. The added value of surgery is uncertain in patients who reach pCR; however, current imaging and biopsy techniques aiming to predict pCR are not accurate enough. This study aims to quantify the residual disease remaining after NST in patients with a favorable response on MRI and residual disease missed with biopsies. Methods In the MICRA trial, patients with a favorable response to NST on MRI underwent ultrasound-guided post-NST 14G biopsies followed by surgery. We analyzed pathology reports of the biopsies and the surgical specimens. Primary outcome was the extent of residual invasive disease among molecular subtypes, and secondary outcome was the extent of missed residual invasive disease. Results We included 167 patients. Surgical specimen showed residual invasive disease in 69 (41%) patients. The median size of residual invasive disease was 18 mm (interquartile range [IQR] 12–30) in hormone receptor-positive (HR+)/human epidermal growth factor receptor 2-negative (HER2−) patients, 8 mm (IQR 3–15) in HR+/HER2-positive (HER2+) patients, 4 mm (IQR 2–9) in HR-negative (HR−)/HER2+ patients, and 5 mm (IQR 2–11) in triple-negative (TN) patients. Residual invasive disease was missed in all subtypes varying from 4 to 7 mm. Conclusion Although the extent of residual invasive disease is small in TN and HER2+ subtypes, substantial residual invasive disease is left behind in all subtypes with 14G biopsies. This may hamper local control and limits adjuvant systemic treatment options. Therefore, surgical excision remains obligatory until accuracy of imaging and biopsy techniques improve.

Clinical problem. Ductal Carcinoma in Situ (DCIS) is a potential precursor for invasive breast cancer (IBC). Therefore, DCIS is currently treated with surgical excision, often supplemented with radiotherapy to prevent progression to ipsilateral IBC (iIBC). However, many DCIS lesions will never do so. Estimating the risk of progression is a grand challenge, as neither the histopathological grade of the DCIS lesion nor other biological markers are conclusively associated with the disease outcome. Aim. We aimed to develop a deep-learning based pipeline for estimating the risk of iIBC recurrence following DCIS using a dataset of 235 H&E-stained whole-slide images (WSIs) of the primary DCIS lesions with corresponding 10-year follow up metadata of DCIS recurrence collected at the Netherlands Cancer Institute. The patients included in our dataset did not receive radiotherapy and experienced recurrence in 167 of the cases. Results. We developed a two-step pipeline that is able to find predictive features for 10-year iIBC prediction. First, tissue regions on WSIs were divided into equally sized tiles. Mammary ducts were detected in the tiles using a RetinaNet with ResNet101 backbone that was implemented in Detectron2 and pre-trained on ImageNet. Selecting only tiles containing ducts served to reduce the input dimensionality of typically giga-pixel WSIs for the second step. Here, DCIS recurrence was predicted by a weakly-supervised multi-instance learning (MIL) classification model where the label of the WSI was determined by average weighting of duct labels. The performance of this model was enhanced by pre-training it with SimCLR, a self-supervised learning method, on image data from the histopathology domain. Our proposed model achieved an AUC of .93 ± .005, with a sensitivity of .83 ± .27 and a specificity of .85± .09. These results show that the model was able to correctly distinguish patients with low subsequent IBC-risk from those with a substantially higher risk. An active research pursuit of our group is now to develop a model which is able to predict iIBC progression for patients treated with radiotherapy. This poses additional challenges to the model, as the effect of radiotherapy, as well as disease outcome must be predicted. Conclusion and impact. Our method opens up an avenue for identifying biologically relevant features for estimating DCIS progression risk into invasive breast cancer. This knowledge may be used for appropriately choosing a personalized DCIS management option for patients - which may be active surveillance rather than surgical removal of the lesion. JW and JT were equal senior co-authors on this project. This work was supported by Cancer Research UK and by KWF Dutch Cancer Society (ref.C38317/A24043) Citation Format: Shannon Doyle, Francesco Dal Canton, Timo Koostra, Maartje van Seijen, Emilie Groen, Efstratios Gavves, Hugo Horlings, Esther Lips, Jelle Wesseling, Jonas Teuwen, Grand Challenge PRECISION Consortium. Deep learning applied on resection specimen tissue slides of ‘pure’ ductal carcinoma in situ predicts ipsilateral invasive breast cancer recurrence [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr PD11-03.

Background: Ductal carcinoma in situ (DCIS) is a potential precursor of invasive breast cancer (IBC). However, the majority of DCIS will never progress to IBC if left untreated. As almost all DCIS are treated by surgery often supplemented by radiotherapy, overtreatment of indolent DCIS occurs. Consequently, there is an urgent need to identify biomarkers to distinguish harmless from aggressive DCIS. We hypothesised that morphometric features observed in Hematoxylin-Eosin (HE) sections of DCIS, such as number, geometry and density of DCIS ducts and cells could identify harmless DCIS lesions.

Background The added value of surgery in breast cancer patients with pathological complete response (pCR) after neoadjuvant systemic therapy (NST) is uncertain. The accuracy of imaging identifying pCR for omission of surgery, however, is insufficient. We investigated the accuracy of ultrasound-guided biopsies identifying breast pCR (ypT0) after NST in patients with radiological partial (rPR) or complete response (rCR) on MRI. Methods We performed a multicenter, prospective single-arm study in three Dutch hospitals. Patients with T1–4(N0 or N +) breast cancer with MRI rPR and enhancement ≤ 2.0 cm or MRI rCR after NST were enrolled. Eight ultrasound-guided 14-G core biopsies were obtained in the operating room before surgery close to the marker placed centrally in the tumor area at diagnosis (no attempt was made to remove the marker), and compared with the surgical specimen of the breast. Primary outcome was the false-negative rate (FNR). Results Between April 2016 and June 2019, 202 patients fulfilled eligibility criteria. Pre-surgical biopsies were obtained in 167 patients, of whom 136 had rCR and 31 had rPR on MRI. Forty-three (26%) tumors were hormone receptor (HR)-positive/HER2-negative, 64 (38%) were HER2-positive, and 60 (36%) were triple-negative. Eighty-nine patients had pCR (53%; 95% CI 45–61) and 78 had residual disease. Biopsies were false-negative in 29 (37%; 95% CI 27–49) of 78 patients. The multivariable associated with false-negative biopsies was rCR (FNR 47%; OR 9.81, 95% CI 1.72–55.89; p = 0.01); a trend was observed for HR-negative tumors (FNR 71% in HER2-positive and 55% in triple-negative tumors; OR 4.55, 95% CI 0.95–21.73; p = 0.058) and smaller pathological lesions (6 mm vs 15 mm; OR 0.93, 95% CI 0.87–1.00; p = 0.051). Conclusion The MICRA trial showed that ultrasound-guided core biopsies are not accurate enough to identify breast pCR in patients with good response on MRI after NST. Therefore, breast surgery cannot safely be omitted relying on the results of core biopsies in these patients.

Background. Ductal carcinoma in situ (DCIS) is a potential precursor of invasive breast cancer (IBC). However, the natural course of a particular DCIS lesion is unknown, because almost all women with DCIS are treated. Furthermore, most studies are biased because these comprise DCIS adjacent to IBC, also known as synchronous DCIS and IBC, indicating that such DCIS lesions already have the capacity to progress to IBC. It is still unknown which proportion and type of subsequent ipsilateral IBCs (iIBC) are related to the initial primary "pure" DCIS lesion. Therefore, we performed an extensive molecular characterization of DCIS and matched subsequent iIBC, to better understand the natural course of DCIS.Patients and methods. We used a unique series of 78 women diagnosed with DCIS and treated by breast conserving surgery (BCS) alone, which subsequently developed iIBC. Mean time to iIBC event was 6.3 years (range 0.5-17.0). These 78 women are a representative sample of a case-control series, nested in a nation-wide, population-based cohort including all patients diagnosed with DCIS between 1989 and 2005 in the Netherlands (Visser, et al Clin Can Res 2018). Data on tumor location (ICD-10) was available for all lesions. DNA and RNA was simultaneously extracted for 78 DCIS lesions and 78 matched subsequent iIBC (DNA >20ng; RNA >100ng), and RNA sequencing (RNAseq) and low coverage whole genome sequencing (CNVseq) was performed. Panel sequencing (PanelSeq), using a custom panel of 53 breast cancer driver genes, was performed with the remaining DNA of 42 DCIS and iIBC matched pairs. We determined if the iIBC lesion and DCIS lesion were related, by comparing tumor location and genomic features.Results. Based on tumor location and histological grade, >95% of the subsequent iIBC reflected outgrowth of residual disease. Based on RNAseq data, 77% of all DCIS and IBC lesions classified into the same PAM50 subtypes. The CNVseq data showed that the DCIS lesions contained copy number aberrations on typical breast cancer-associated loci, such as 1q gain, 8q gain, 16q loss, 20q gain. However, when we compared DCIS with their matched iIBC, we observed in 41% of the cases very distinct copy number profiles, indicating either outgrow of a different tumor (minority) subclone or a second primary tumor. Analysis of PanelSeq mutation data supported this clonal or independent origin of the subsequent iIBC.Conclusion. To our knowledge, our study is the first to investigate if subsequent iIBC is likely to originate from the initial primary "pure" DCIS in a large series with long-term follow-up.Surprisingly, our CNVseq and PanelSeq results indicate that more than one third of the subsequent iIBCs after BCS alone treated primary DCIS are likely to be second primary tumors or represent selective outgrowth of a minority DCIS subclone.Citation Format: Lindy Visser, Marlous Hoogstraat, Tycho Bismeijer, Lotte Elshof, Koen van de Vijver, Emilie Groen, Mathilde Almekinders, Joyce Sanders, Carolien Bierman, Dennis Peters, Ingrid Hofland, Frank Nieboer, Michiel de Maaker, Petra Kristel, Lennart Mulder, Annegien Broeks, Michael Schaapveld, Marjanka Schmidt, Lodewyk Wessels, Esther Lips, Jelle Wesseling, On behalf of the PRECISION team. Ductal carcinoma in situ of the breast: Cancer precursor or not [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 751.

535 Background: Young age at breast cancer diagnosis is considered a poor prognostic factor. As a result, many treatment guidelines advice adjuvant systemic treatment for young patients. Answering prognostic questions on young patients has therefore become a challenge. The PARADIGM (PAtients with bReast cAncer DIaGnosed preMenopausally) project aims to assess the long-term outcome of women diagnosed with breast cancer ≤40 years in the absence of adjuvant systemic therapy, using real world data from the nationwide Netherlands Cancer Registry (NCR) coupled with tissue biobanking. Methods: All women ≤40 years, diagnosed in the Netherlands between 1989-2000 with a primary invasive, histologically proven, TanyN0M0 breast cancer, without adjuvant systemic treatment were identified through the NCR. Back then N0 patients were considered low risk and did not receive adjuvant systemic treatment. Tissue specimens were revised by a team of dedicated breast pathologists. Cox regression was performed to estimate hazard ratios for recurrence-free (RFS) and overall survival (OS) according to immunohistochemical (IHC) subtype. Analyses were adjusted for grade, pathological T-stage, histological subtype and radiotherapy. Results: We included 2310 patients with a mean follow-up of 15.4 years (range 0-25 years). OS for the whole cohort was 68% and RFS 58.4% at 25 years. In total 740 deaths and 1043 recurrences were observed. Hormone receptor (HR)+/HER2+ patients had a significantly worse OS when compared to HR-HER2+ patients (adjusted Hazard Ratio 1.58; 95% confidence interval 1.05-2.38; p=0.029). No difference was observed between HR-HER2+ and the triple negative and HR+/HER2- subgroups at 25-years. RFS was similar for all IHC subtypes. Conclusions: In this large cohort of non-adjuvant systemically treated young breast cancer patients with long-term follow-up HR+/HER2+ patients have a significantly worse survival when compared to triple negative, HR-/HER2+ and HR+/HER2- patients. The latter three subtypes have similar OS at 25 years. Future molecular studies have been planned to distinguish the favorable from the unfavorable prognostic patients.

Neoadjuvant treatment of HER2-positive breast cancer frequently leads to a pathologic complete response (pCR), which is associated with favourable long-term outcome. Treatment regimens typically consist of 6–9 cycles of trastuzumab-based chemotherapy, although many patients achieve early radiologic complete response (rCR). If rCR accurately predicts pCR, the number of chemotherapy cycles can possibly be reduced.

INTRODUCTION:Currently used tools for breast cancer prognostication and prediction may not adequately reflect a young patient's prognosis or likely treatment benefit because they were not adequately validated in young patients. Since breast cancers diagnosed at a young age are considered prognostically unfavourable, many treatment guidelines recommend adjuvant systemic treatment for all young patients. Patients cured by locoregional treatment alone are, therefore, overtreated. Lack of prognosticators for young breast cancer patients represents an unmet medical need and has led to the initiation of the PAtients with bReAst cancer DIaGnosed preMenopausally (PARADIGM) initiative. Our aim is to reduce overtreatment of women diagnosed with breast cancer aged ≤40 years. METHODS AND ANALYSIS:All young, adjuvant systemic treatment naive breast cancer patients, who had no prior malignancy and were diagnosed between 1989 and 2000, were identified using the population based Netherlands Cancer Registry (n=3525). Archival tumour tissues were retrieved through linkage with the Dutch nationwide pathology registry. Tissue slides will be digitalised and placed on an online image database platform for clinicopathological revision by an international team of breast pathologists. Immunohistochemical subtype will be assessed using tissue microarrays. Tumour RNA will be isolated and subjected to next-generation sequencing. Differences in gene expression found between patients with a favourable and those with a less favourable prognosis will be used to establish a prognostic classifier, using the triple negative patients as proof of principle. ETHICS AND DISSEMINATION:Observational data from the Netherlands Cancer Registry and left over archival patient material are used. Therefore, the Dutch law on Research Involving Human Subjects Act (WMO) is not applicable. The PARADIGM study received a 'non-WMO' declaration from the Medical Ethics Committee of the Netherlands Cancer Institute - Antoni van Leeuwenhoek hospital, waiving individual patient consent. All data and material used are stored in a coded way. Study results will be presented at international (breast cancer) conferences and published in peer-reviewed, open-access journals.

Abstract Background A study by Cheang et al. reported that for ER-positive, HER2-negative breast cancer a Ki67 score ≥14% distinguishes Luminal B from Luminal A tumors. The (neo)adjuvant treatment of these Luminal B tumors consists of endocrine treatment and chemotherapy while chemotherapy can be omitted for Luminal A tumors. For the determination of these subtypes, resection specimens and particularly core biopsies in the neoadjuvant setting are used. Since Ki67 scoring is advocated by some to be a marker in deciding if systemic chemotherapy will be given or not, its assessment should be highly reproducible and reliable. I.e, Ki67 assessment should not vary significantly between different laboratories, but also not within the tumor. To test this, we analyzed (1) interlaboratory agreement using whole slides and (2) agreement within a particular tumor using tissue microarrays (TMA) as well as whole slides from surgical specimens. Material and methods The first 100 patients from the microarRAy prognoSTics-in-breast-cancER (RASTER) study (n=427) were selected. From each patient clinicopathological characteristics and tumor blocks were available. From each tumor block 2 whole slides were cut for staining in laboratory 1 and 2. Also six single cores were taken to construct 2 tissue microarrays (core 1-3 and core 4-6). Intrinsic subtypes were defined as follows: Luminal A, Luminal B/HER2-, Luminal B/HER2+, HER2-overexpressing, and Basal-like. Experienced pathologists (EG, JS, VTHMBS, JW) performed the scoring. Results There were 99 whole slides suitable for analysis. Substantial agreement (κ = 0.723, P<0.001) was demonstrated in discrimination for Ki67 low and high between whole slides stained in laboratory 1 and laboratory 2 with a discordance rate of 13.1% (Table 1). Moderate agreement (κ = 0.595, κ = 0.584, P<0.001) was demonstrated in discrimination of Ki67 low and high between whole slide and core 1-3 and core 4-6 with a discordance rate of 20.3% and 20.8% respectively (Table 2). Substantial agreement (κ = 0.666) was demonstrated for Ki67 expression between core 1-3 and core 4-6. Table 1. Discordance between laboratories for whole slide Ki67 result Laboratory 2 <14%≥14%Discordance rate %KappaSign.Laboratory 1<14%56013.10.723<0.001 ≥14%1330 Table 2. Discordance between whole slide and core biopsies 1-3 and 4-6 for Ki67 result Whole slide <14%≥14%Discordance rate %KappaSign.Core 1-3<14%33920.30.595<0.001 ≥14%730 Core 4-6<14%31820.80.584<0.001 ≥14%830 Conclusion Tumor heterogeneity results in substantial variation of Ki67 scores between TMA cores and whole slides that may result in considerable differences in distinguishing luminal A from luminal B ER-positive, HER2-negative tumors. This may have far reaching consequences for the choice for (neo)adjuvant treatment. Citation Format: Matthijs V Nijenhuis, Emilie Groen, Tim J Dekker, Caroline A Drukker, J Sanders, V T Smit, S Linn, E J Rutgers, J Wesseling. Tumor heterogeneity impairs robustness of Ki67 scoring in breast cancer [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P5-02-04.

I.J. Clarke, Clayton, Vic. C. Coen, London M.A. Cowley, Clayton, Vic. W.W. de Herder, Rotterdam S.L. Dickson, Gothenburg A. Enjalbert, Marseille A.B. Grossman, Headington D. Jezová, Bratislava A.S. Kauff man, La Jolla, Calif. M.J. Kelly, Portland, Oreg. P.A. Kelly, Paris X. Ni, Shanghai S.R. Ojeda, Beaverton, Oreg. (Reviews) G. Rindi, Rome H. Sasano, Sendai J.Y. Seong, Seoul M. Tena-Sempere, Cordoba H. Tsukamura, Nagoya E.J. Wagner, Pomona, Calif. B. Wiedenmann, Berlin Z. Zeng, Beijing Editorial Board

Background: The latest WHO classification for neuroendocrine neoplasms (NEN) of the gastrointestinal tract defines grade according to Ki-67 and mitotic indices. Some have questioned the reproducibility and thus the reliability of Ki-67 assessment. We therefore investigated the accuracy of this proliferation marker in NEN. Methods: The Ki-67 index of tumor specimens of NEN (n = 73) was assessed by two pathologists as in routine practice with eyeballing and twice by image analysis using ImageJ freeware at different magnifications. Results were correlated with overall survival. Results: The intraclass correlation coefficient (ICC) between pathologists was 0.88. The ICC for the measurements using image analysis was 0.85. The ICC between all four measurements (pathologists and ImageJ) was 0.80. If the Ki-67 index was translated to grade as prescribed by the current WHO classification (<3% = grade 1, 3-20% = grade 2, >20% = grade 3), kappa was between 0.61 and 0.75. Grades based on pathologist scoring were often (16-29%) higher than grades assigned by image analysis (p < 0.001). Grade was significantly correlated with survival (p < 0.0001) irrespective of the way Ki-67 was assessed. Conclusion: Assessment of the Ki-67 index by eyeballing correlates remarkably well with the Ki-67 index as calculated by image analysis and is therefore an accurate parameter. Moreover, it is significantly related to survival irrespective of the method used. Yet if the Ki-67 index is translated to grade, the grade should be interpreted with caution due to values around threshold levels.