ECP 2023 Abstracts

S62 Virchows Archiv (2023) 483 (Suppl 1):S1–S391 13 self-assessed by each participant then anonymized for assessment by the CPQA-AQCP in direct comparison to results obtained from a laboratory with validated testing for detection of HER2-low expression. Results: This pilot scheme for HER2 immunostaining in HER2-low breast cancers ran in November 2022, with 35 participating laboratories. Two general observations will be discussed: 1) many participants applied a very conservative approach to diagnosis of 1+ HER2 immunostaining in their self-assessment that was corrected upon review by the CPQA-AQCP, and 2) the overall technical quality of staining was generally sufficient to detect HER2-low expressing breast cancers as defined by reference labora- tory staining. Lack of sensitivity in detection of low expression was not a problem for most laboratories if cores were interpreted correctly. Detection of additional HER2-low expressing tumours compared to the reference laboratory led to poor specificity for many participants. Conclusion: As the treatment landscape has changed, so too has assessment of relevant biomarkers. Reporting of HER2 immunostain- ing must include the IHC score (0, 1+, 2+, 3+). Laboratories and pathologists may wish to reconsider their reporting terminology and specifically indicate “HER2-low” for cases showing IHC 1+ or IHC 2+ not amplified by ISH. At this early point in validation of immunostain- ing to detect HER2-low expressing breast cancers, continued participa- tion in external quality assurance and proficiency testing will be vital. Funding: AstraZeneca PS-01-022 MHC class I and PD-L1 expression in ductal carcinoma in situ sub- types of the breast: impact on T-cell infiltration and clinical outcome J.S. Lee*, M.H. Park, N.I. Kim *Department of Pathology, Chonnam National University Hwasun Hospital, Republic of Korea Background & objectives: PD-L1 expression and MHC class I down- regulation are essential mechanisms of tumour escape from the host’s immune system. We examined MHC class I expression in ductal car- cinoma in situ (DCIS) subtypes with attention to PD-L1 expression and T cell infiltration. Methods: MHC class I and PD-L1 expression and CD3+ and CD8+ T lymphocytes were detected with immunohistochemistry in 131 pure DCIS samples using tissue microarrays. DCISs were classified into hormone receptor (HR)+/ human epidermal growth factor 2 (HER2)-, HR+/HER2+, HR-/HER2+, and triple-negative (TN) subtype. Results: Loss of MHC class I expression was found in 16.4% (21/128) cases and was highest in the HR+/HER2- subtype (29.2%). PD-L1 expression in immune cells (≥1%) was seen in 18.8% (24/128). PD-L1 expression was higher in HER2+ (HR+/HER2+ and HR-/HER2+) subtypes than in HR+/HER2- subtype. In PD-L1 positive DCIS, MHC class I expression was intact in most cases (95.8%), whereas only one case (4.2%) lost MHC I expression in HR-/HER2+ subtype. In HR+/ HER2- and HR-/HER2+ subtypes, the numbers of stromal CD3+ and CD+8 T cells was highest in MHC class I intact and PD-L1 positive cases. MHC class I expression and PD-L1 expression were not associ- ated with tumour recurrence. Conclusion: We demonstrated differential patterns of MHC class I and PD-L1 expression and T-cell infiltration in distinct subtype of DCIS. Increased knowledge regarding MHC class I, PD-L1, and immune sub- set including CD8+ T cells could contribute to future development of immune modulating therapies in DCIS, especially HER2+ subtype. PS-01-023 Impact of HER2 assay sensitivity on the ability to detect tumours with low to ultra-low HER2 expression: a TMA study on more than 10,000 tumours from more than 100 tumour entities M. Lennartz*, N. Blessin, D. Dum, D. Hoeflmayer, S. Weidemann, G. Sauter, T.S. Clauditz, F. Jacobsen, T. Krech, A.H. Marx, S. Minner, R. Simon, N. Gorbokon, S. Steurer, E. Burandt *University Medical Center Hamburg-Eppendorf, Germany Background & objectives: Detection of low level HER2 expression is important for therapy with new HER2 inhibitors in breast cancer. HER2 assays with high sensitivity are therefore needed. These may identify low level HER2 expression in various types other than breast cancer. Methods: A tissue microarray containing >10,000 tissue samples from more than 100 different tumour types and subtypes was ana- lysed for immunohistochemical expression of HER2 using different immunohistochemistry protocols and anti-HER2 antibodies, includ- ing the HercepTest and laboratory developed tests designed for high sensitivity. HER2 IHC evaluation included the recording of HER2 low (1+), ultralow (any staining, less than 1+), and 0 (negative). Results: Irrespective of the assay sensitivity all assays showed the expected/tolerable association with HER2 FISH data in 308 breast cancers. In non-breast cancers, the use of different assays yielded var- iable rates of low or ultra-low HER2 expression in different tumour entities. For example, low (1+) and ultra-low (+) expression was observed in 7-24% (1+) and 11-28% (+) of serous ovarian cancers, 6.2-24% (1+) and 3.4-19% (+) of endometrioid endometrial cancers, 5.5-14% (1+) and 7.7-13% (+) of intestinal gastric cancers, 1.8-9.2% (1+) and 4.5-11% (+) of ductal adenocarcinomas of the pancreas, 19-33% (1+) and 5.5-11% (+) of muscle invasive urinary bladder can- cers, 2.9-16% (1+) and 4.3-12% (+) of adenocarcinomas of the colon. Conclusion: Low level HER2 expression occurs commonly in many cancer types. The rate of identifiable tumours with HER2 low or HER2 ultra-low expression varies depending on assay parameters that can easily be modified. Irrespective of assay sensitivity, our data identify various tumour entities with significant fractions of “HER2- low” cases that might benefit from therapy with new antibody drug conjugates. PS-01-024 Discordance in molecular (Prosigna ®) and surrogate (immuno- histochemistry) subtyping of breast carcinoma G.A. Martín Small*, A. Castillo, L. Pons, C. Perelló-Fabregat, P. Rodríguez-Martínez, E. Castellà, A.M. Muñoz-Mármol, A. Urbizu, L. Arnaldo, M. Martín-Céspedes, R. Marginet-Flinch, M. Jimeno, M. Margelí, C. Sanz, P.L. Fernández *Hospital Universitari Germans Trias i Pujol, Spain Background & objectives: Prosigna is a genetic test that analyses the expression of 50 genes estimating a 10-year risk of relapse classifying the tumours in one molecular subtype of breast cancer (BC). This classification doesn’t always correlate with the immunohistochemistry (IHC) surrogate subtype. Methods: 199 invasive BC were studied with Prosigna. They had been classified as Luminal A (LA), Luminal B/Her2- (LB/HER2-), LB/ HER2+, HER2+ and Triple Negative (TN) by IHC (St Gallen 2021 guidelines). A Ki67 cut-off point of 24% (institutional mean) and 20% for progesterone receptors (PR) was used to categorized luminal subtypes. Agreement between both methods was estimated with kappa coefficient. Results: There were 77 discordant cases (38.7%). Most of them (n=39) were tumours classified as LA by Prosigna but as LB/HER2- by IHC due to low progesterone receptors (PR) (n=20), high Ki67 index (n=16) or both (n=3). The majority of the LB cases correctly classi- fied had a Ki67≥24%. We also identified 20 LB cases (Prosigna), classified as LA by IHC. Concordance rate between both methods was poor in Luminal subtypes (Cohen’s kappa, k=0,32). Moreover, 7 of the discordant cases classified as HER2enriched by Prosigna were LB/HER2- (4), LB/HER2+ (1) and TN (2) by IHC. This may be due to HER2 activating mutation that doesn’t overexpress HER2 protein.