Immunohistochemical Algorithms Research Articles

Improving histotyping precision: The impact of immunohistochemical algorithms on epithelial ovarian cancer classification

To improve the precision of epithelial ovarian cancer histotyping, Köbel et al. (2016) developed immunohistochemical decision-tree algorithms. These included a six- and four-split algorithm, and separate six-split algorithms for early- and advanced stage disease. In this study, we evaluated the efficacy of these algorithms. A gynecological pathologist determined the hematoxylin and eosin (H&E)-based histotypes of 230 patients. Subsequently, the final histotypes were established by re-evaluating the H&E-stained sections and immunohistochemistry outcomes. For histotype prediction using the algorithms, the immunohistochemical markers Napsin A, p16, p53, progesterone receptor (PR), trefoil factor 3 (TFF3), and Wilms’ tumor 1 (WT1) were scored. The algorithmic predictions were compared with the final histotypes to assess their precision, for which the early- and advanced stage algorithms were assessed together as six-split-stages algorithm. The six-split algorithm demonstrated 96.1% precision, whereas the six-split-stages and four-split algorithms showed 93.5% precision. Of the 230 cases, 16 (7%) showed discordant original and final diagnoses; the algorithms concurred with the final diagnosis in 14/16 cases (87.5%). In 12.4%–13.3% of cases, the H&E-based histotype changed based on the algorithmic outcome. The six-split stages algorithm had a lower sensitivity for low-grade serous carcinoma (80% versus 100%), while the four-split stages algorithm showed reduced sensitivity for endometrioid carcinoma (78% versus 92.7–97.6%). Considering the higher sensitivity of the six-split algorithm for endometrioid and low-grade serous carcinoma compared with the four-split and six-split-stages algorithms, respectively, we recommend the adoption of the six-split algorithm for histotyping epithelial ovarian cancer in clinical practice.

Correlation between Germinal Center Differentiation and Double/Triple Hit Score in the Classification of Diffuse Large B Cell Non-Hodgkin Lymphoma: A Clinicopathological Approach

Background: Diffuse large B cell lymphoma (DLBCL) is considered the most common type of non-Hodgkin lymphoma. The global incidence of DLBCL has been doubled in the past decades, highlighting the need for more effective treatment regimens. The curability of DLBCL is heavily influenced by a number of factors, such as the age, the international prognostic index (IPI) score, the molecular cell of origin (COO) subtype, and presence/ absence of specific chromosomal rearrangements or protein expression. Recent gene profiling studies have classified DLBCL into two main categories, germinal center B-cell type (GCB) and activated B-cell type (ABC). Validation of these subtypes has become more applicable after the immunohistochemical algorithm suggested by Hans, which included three commercially available markers, CD10, BCL6, and MUM-1/IRF4. In addition to this classification, DLBCL prognostic models based on different sets of genes or immunohistochemical markers have been proposed. More recently, DLBCLs with translocations of MYC, along with a B-cell lymphoma 2 (BCL2) and/or B-cell lymphoma 6 (BCL6) rearrangement, are now called double-hit lymphoma (DHL) or triple-hit lymphoma (THL), respectively. Furthermore, the co-expre sion of MYC and BCL2 proteins without underlying rearrangements is considered a new adverse prognostic indicator termed double-expressor lymphoma (DEL). The present study is aiming to combine morphological, immunophenotypical, and genetic features of DLBCLs to reach a more reproducible subtyping. Aim: To identify the relation between clinicopathological features as patient age, sex, site of tumor, clinical presentation and tumor stage and the germinal center differentiation in DLBCL (germinal center B-cell (GCB) or activated B-cell (ABC)) and the double/triple genetic hits status.Material and Methods: The material of this study comprised 52 formalin-fixed paraffin- embedded blocks representative of diffuse large B cell lymphoma cases retrieved from the archival material available at the Pathology Department, Faculty of Medicine, Alexandria University. Subclassification of cases into (germinal center / and post-germinal center types) using a panel of immunohistochemical markers including CD10, BCL6 and MUM1 was done and scoring of double/triple hit expression using a panel of immunohistochemical markers including c-MYC, BCL6 and BCL2 was also done and correlation with their subtypes and clinical data was performed.Results: there was significant association between germinal center differentiation and tumor stage (MCp= 0.018).Conclusion: Germinal center differentiation status is a valid method to evaluate the prognosis of DLBCL.

Molecular-Based Immunohistochemical Algorithm for Uterine Leiomyosarcoma Diagnosis

The morphologic assessment of uterine leiomyosarcoma (LMS) may be challenging, and diagnostic immunohistochemical (IHC) analysis is currently lacking. We evaluated the genomic landscape of 167 uterine LMS by targeted next-generation sequencing (NGS) to identify common genomic alterations. IHC analyses corresponding to these genomic landmarks were applied to a test cohort of 16 uterine LMS, 6 smooth muscle tumors of uncertain malignant potential (STUMP), and 6 leiomyomas with NGS data and a validation cohort of 8 uterine LMS, 12 STUMP, 21 leiomyomas and leiomyoma variants, 7 low-grade endometrial stromal sarcomas, and 2 diagnostically challenging uterine smooth muscle tumors. IHC results were individually interpreted by 3 pathologists blinded to NGS data. Overall, 94% of LMS showed ≥1 genomic alteration involving TP53, RB1, ATRX, PTEN, CDKN2A, or MDM2, with 80% showing alterations in ≥2 of these genes. In the test cohort, an initial panel of p53, Rb, PTEN, and ATRX was applied, followed by a panel of DAXX, MTAP, and MDM2 in cases without abnormalities. Abnormal p53, Rb, PTEN, and ATRX IHC expression was seen in 75%, 88%, 44%, and 38% of LMS, respectively, in the test cohort. Two or more abnormal IHC results among these markers were seen in 81% of LMS. STUMPs demonstrated only 1 IHC abnormality involving these markers. No IHC abnormalities were seen in leiomyomas. In the validation cohort, abnormal p53, Rb, and PTEN IHC results were seen in LMS, whereas rare STUMP or leiomyomas with bizarre nuclei showed IHC abnormalities involving only 1 of the markers. Abnormalities in ≥2 markers were present in both diagnostically challenging smooth muscle tumors, confirming LMS. Concordance was excellent among pathologists in the interpretation of IHC (κ = 0.97) and between IHC and NGS results (κ = 0.941). Uterine LMS exhibit genomic landmark alterations for which IHC surrogates exist, and a diagnostic algorithm involving molecular-based IHC may aid in the evaluation of unusual uterine smooth muscle tumors.

Primary central nervous system lymphoma: Comprehension of cell-of-origin subtypes.

To classify PCNS-DLBCL into COO-subtypes based on immunohistochemical algorithms by Hans and Choi and evaluate concordance between the two. A further aim is to investigate the clinicoradiological and histomorphological parameters of the subtypes thus obtained. As many as 143 cases of primary CNS lymphoma were evaluated by immunohistochemistry for CD10, BCL6, MUM1, GCET, and FOXP1 and based on which the said 143 cases were further classified into COO subtypes using Hans and Choi algorithms. Mean age was 53.8 years with marginal male preponderance and predominantly centroblastic morphology (75.5%). CD 10 was positive in 8.9% of the cases, BCL6 in 58.6%, MUM1 in 89.9%, GCET in 32.9%, and FOXP1 in 79.5%. As much as 84.9% cases were of non-germinal center B-cell (GCB) subtype and 15.1% cases were of GCB subtype as determined based on Hans algorithm. Furthermore, 90.7% cases were of activated B-cell (ABC) subtype and 9.3% cases were of GCB subtype according to Choi algorithm. A 91.8% concordance was observed between Hans and Choi algorithms. Among the 6 discordant cases, 5 cases were subtyped as GCB by Hans and ABC by Choi and 1 case as ABC by Hans and GCB by Choi. Most of PCNS-DLBCLs are of non-GCB/ABC COO subtype, but inconsistences abound in the utility of IHC algorithms in PCNS-DLBCL COO subtypes.

Immunohistochemical Algorithm for the Classification of Muscle-Invasive Urinary Bladder Carcinoma with Lymph Node Metastasis: An Institutional Study.

Muscle-invasive urothelial carcinoma represents 20% of newly diagnosed cases of bladder cancer, and most cases show aggressive biological behavior with a poor prognosis. It is necessary to identify biomarkers that can be used as prognostic and predictive factors in daily clinical practice. In our study, we analyzed different antibodies in selected cases of muscle-invasive urinary bladder carcinoma and lymph node metastasis to identify immunohistochemical types and their value as possible prognostic factors. A total of 38 patients were included, 87% men and 13% women, with a mean age of 67.8 years. The most frequent histopathological type was urothelial carcinoma. In the primary lesion, the mixed type was the most common. In unilateral metastasis, the mixed type was the most frequently found. In cases of primary lesions and bilateral metastasis, the luminal and mixed types were observed. The luminal subtype was the most stable in immunohistochemical expression across primary tumors and metastases. The basal type showed a better prognosis in terms of disease-free survival. In conclusion, immunohistochemical studies are useful in assessing primary and metastatic lesions in patients with urothelial carcinoma. Immunohistochemical classification can typify muscle-invasive urothelial carcinoma, and the immunophenotype seems to have prognostic implications.

Immunohistochemistry subtyping of urothelial carcinoma is feasible in the daily practice

The preferred treatment of choice in muscle-invasive bladder cancer (MIBC) is usually transurethral resection followed by cystectomy, with neoadjuvant chemotherapy being a second option. As the treatment is associated with relevant side effects, a great effort is being made to improve the selection of patients, with molecular subtyping being one of the main strategies. Our aim was to develop an immunohistochemical algorithm for subtyping MIBCs. After a literature review, we have developed a simple algorithm to subtype MIBCs based on their morphology and three common antibodies: GATA3, CK5/6, and p16. We applied it to 113 muscle-invasive carcinomas. The positivity threshold for GATA3 and CK5/6 was 20% with at least moderate intensity, while p16 was 70% with moderate to intense nuclear and cytoplasmic staining. Cases GATA3 + CK5/6 - were considered luminal, while cases GATA3 - CK5/6 + were classified as nonluminal/basal squamous. Luminal p16 + cases were labeled as genomically unstable and luminal p16 - as Uro-like. Cases GATA3 + CK5/6 + with a predominantly basal pattern were labeled luminal, while diffuse cases were labeled nonluminal/basal squamous. All GATA3-CK5/6 - cases were considered nonluminal and were divided into mesenchymal-like or neuroendocrine, depending on the morphology. We were able to classify the 113 cases as: 82 (72.57%) were luminal, being 47 Uro-like (41.59%) and 35 (30.97%) genomically unstable; 31 (27.43%) were nonluminal, being 24 basal/squamous (21.24%), two (1.76%) mesenchymal-like, and five (4.42%) neuroendocrine like. We have achieved a feasible and cost-effective algorithm to subtype MIBCs from morphological features and the use of three common antibodies. Further studies in external cohorts are necessary to validate these results.

The role of cytokeratin 7/20 coordination revisited-Machine learning identifies improved interpretative algorithms for cell block immunohistochemistry in aspirates of metastatic carcinoma.

Fine-needle aspiration (FNA) is a robust diagnostic technique often used for tissue diagnosis of metastatic carcinoma. For interpretation of FNA cytology, cell block immunohistochemistry (IHC) and clinicocytologic parameters are indispensable. In this review of a large cohort, the current report: 1) describes clinicocytologic parameters and immunoprofiles of aspirates of metastatic carcinoma, 2) compares the predictivity of immunostains and classical approaches for IHC interpretation, and 3) describes machine learning-based algorithms for IHC interpretation. Aspirates of metastatic carcinoma that had IHC performed were retrieved. Clinicocytologic parameters, IHC results, the corresponding primary site, and histologic diagnoses were recorded. By using machine learning, decision trees for predicting the primary site were generated, their performance was compared with 2 human-designed algorithms, and the primary site was suggested in the historical diagnosis. In total, 1145 cases were identified. The 6 most populated groups were selected for machine learning and predictive analysis. With IHC input, the decision tree achieved a concordance rate of 94.5% and overall accuracy of 83.6%, which improved to 95.3% and 85.8%, respectively, when clinical data were incorporated and exceeded the human-designed IHC algorithms (P < .001). The historical diagnosis was more accurate unless indeterminate diagnoses were regarded as discordant (P < .001). CDX2 and TTF-1 immunostains had the highest weight in model accuracy, occupied the root of the decision trees, scored higher as features of importance, and outperformed the predictive power of cytokeratins 7 and 20. Cytokeratins 7 and 20 may be superseded in immunostaining panels, including organ-specific immunostains such as CDX2 and TTF-1. Machine learning generates algorithms that surpasses human-designed algorithms but is inferior to expert assessment integrating clinical and cytologic assessment.

Immunohistochemical algorithms and gene expression profiling in primary cutaneous B-cell lymphoma

Objectiveto assess whether immunohistochemical (IHC) algorithms used to classify the cell of origin (COO) of nodal Diffuse Large B-cell lymphoma (nDLBCL) in Germinal Center type (GCB) and non-GCB subtypes may be applied to Primary Cutaneous B-cell lymphoma (PCBCL) too, and which of these algorithms performs better on PCBCL. DesignRetrospective case control study. SettingPathology Department of the University Hospital “San Giovanni di Dio e Ruggi d′Aragona” Salerno, Italy. ParticipantsFourteen PCBCL, including Primary Cutaneous follicle centre lymphoma (PCFCL) and primary cutaneous diffuse large B-cell lymphoma, Leg type (PCDLBCL-LT) and 14 nDLBCL were evaluated for 7-year period (January 2011 to December 2017). Primary cutaneous marginal zone cell lymphoma (PCMZL) cases were not included in the present study. InterventionEvaluation of immunohistochemical CD10, BCL6, MUM1/IRF4, BCL2, MYC and Ki-67 expression and classification according to three different algorithms. Gene expression profiling (GEP) was performed on the same series using Lymph2Cx assay (Nanostring). The data obtained were compared and analysed. ResultsAll the IHC algorithms showed 13 GCB and 15 non-GCB. GEP showed 12 GCB, 12 activated B cell–type and 4 unclassified. ConclusionsThe PCBCL were classifiable as GCB and non-GCB like the nDLBCL as IHC algorithms were concordant to GEP and produced the same results.

Pheochromocytomas and Abdominal Paragangliomas: A Practical Guidance.

Simple SummaryPheochromocytomas and abdominal paragangliomas (PPGLs) are rare. They can be discovered incidentally by imaging with computed tomography or magnetic resonance imaging and during hormonal surveillance in patients with known genetic variants that are associated with PPGLs. As most PPGLs are functioning, a hormonal work-up evaluating for catecholamine excess is recommended. Classical symptoms, such as tachycardia, hypertension and headache, can be present, but when the PPGL is discovered as an incidentaloma, symptoms may be lacking or be more discrete. PPGLs carry malignant potential, and patients should undergo close surveillance, as recurrence of disease or metastasis may develop. Genetic susceptibility for multifocal disease has gained more attention, and germline variants are commonly detected, thus facilitating detection of hereditary cases and afflicted family members. Any patient with a PPGL should be managed by an expert multidisciplinary team consisting of endocrinologists, radiologists, surgeons, pathologists and clinical geneticists.Pheochromocytomas and abdominal paragangliomas (PPGLs) are rare tumors arising from the adrenal medulla or the sympathetic nervous system. This review presents a practical guidance for clinicians dealing with PPGLs. The incidence of PPGLs has risen. Most cases are detected via imaging and less present with symptoms of catecholamine excess. Most PPGLs secrete catecholamines, with diffuse symptoms. Diagnosis is made by imaging and tests of catecholamines. Localized disease can be cured by surgery. PPGLs are the most heritable of all human tumors, and germline variants are found in approximately 30–50% of cases. Such variants can give information regarding the risk of developing recurrence or metastases as well as the risk of developing other tumors and may identify relatives at risk for disease. All PPGLs harbor malignant potential, and current histological and immunohistochemical algorithms can aid in the identification of indolent vs. aggressive tumors. While most patients with metastatic PPGL have slowly progressive disease, a proportion of patients present with an aggressive course, highlighting the need for more effective therapies in these cases. We conclude that PPGLs are rare but increasing in incidence and management should be guided by a multidisciplinary team.

Analysis of prognostic factors in diffuse large B-cell lymphoma associated with rheumatic diseases

ObjectiveThe risk of developing diffuse large B-cell lymphoma (DLBCL) is increased in many rheumatic diseases (RDs). It is possible that RD-associated DLBCL is a distinct subset within the category of...

Diagnostic and prognostic implications of a three-antibody molecular subtyping algorithm for non-muscle invasive bladder cancer.

Intrinsic molecular subtypes may explain marked variation between bladder cancer patients in prognosis and response to therapy. Complex testing algorithms and little attention to more prevalent, early‐stage (non‐muscle invasive) bladder cancers (NMIBCs) have hindered implementation of subtyping in clinical practice. Here, using a three‐antibody immunohistochemistry (IHC) algorithm, we identify the diagnostic and prognostic associations of well‐validated proteomic features of basal and luminal subtypes in NMIBC. By IHC, we divided 481 NMIBCs into basal (GATA3−/KRT5+) and luminal (GATA3+/KRT5 variable) subtypes. We further divided the luminal subtype into URO (p16 low), URO‐KRT5+ (KRT5+), and genomically unstable (GU) (p16 high) subtypes. Expression thresholds were confirmed using unsupervised hierarchical clustering. Subtypes were correlated with pathology and outcomes. All NMIBC cases clustered into the basal/squamous (basal) or one of the three luminal (URO, URO‐KRT5+, and GU) subtypes. Although uncommon in this NMIBC cohort, basal tumors (3%, n = 16) had dramatically higher grade (100%, n = 16, odds ratio [OR] = 13, relative risk = 3.25) and stage, and rapid progression to muscle invasion (median progression‐free survival = 35.4 months, p = 0.0001). URO, the most common subtype (46%, n = 220), showed rapid recurrence (median recurrence‐free survival [RFS] = 11.5 months, p = 0.039) compared to its GU counterpart (29%, n = 137, median RFS = 16.9 months), even in patients who received intravesical immunotherapy (p = 0.049). URO‐KRT5+ tumors (22%, n = 108) were typically low grade (66%, n = 71, OR = 3.7) and recurred slowly (median RFS = 38.7 months). Therefore, a simple immunohistochemical algorithm can identify clinically relevant molecular subtypes of NMIBC. In routine clinical practice, this three‐antibody algorithm may help clarify diagnostic dilemmas and optimize surveillance and treatment strategies for patients.

A New Immunohistochemical Algorithm Using Four Antibodies Has a High Concordance with Gene Expression Profiling Defined ABC and GCB Diffuse Large B-Cell Lymphoma

▪The new WHO 2016 classification of lymphoid neoplasms requires the identification of Activated B-cell like (ABC) and Germinal center B-cell like (GCB) Diffuse large B-cell lymphomas (DLBCL) NOS as oncogenetic pathways and probably sensitivity to specific therapies differ between these two subtypes. Whereas RNA based signatures from FFPE DLBCL seem promising in reproducibility, still 15% to 20% of DLBCL NOS are of the Unclassified subtype and there is variability between gene expression profiling (GEP) and RNA signatures considering this unclassified subtype. Therefore, the immunohistochemical tool, used during the diagnostic process of the lymphoma, might be useful to maintain and develop in addition to the molecular tools. Hans and coworkers previously developed an immunohistochemical algorithm with 86% concordance with GEP among patients treated with Rituximab with antracyclin based regimens. However, some antibodies such as BCL6 are difficult to evaluate and might hamper reproducibility of scoring. We propose an alternative to the Hans algorithm using 4 easy to use antibodies while keeping the classification process fairly simple.MethodsOne thousand eight hundred forty-three patients with previously untreated de novo CD20+ DLBCL were enrolled in the GELA (Groupe d'Etudes des Lymphomes de l'Adulte)/LYSA (Lymphoma Study Association) LNH01-5B and LNH03-B clinical trials. A Tissue microarray of 854 tumors were built and immunostaining with CD10, MUM1, FOXP1 (Barrans' threshold, 0 vs Variable/Strong), BCL6, IGM, MYC, BCL2 was performed. 761 cases of DLBCL NOS had immunostains data available. For 150 of these patients, frozen tissue was available and GEP has been analyzed using GeneChip Human Genome HGU133 Plus 2.0 arrays (Affymetrix, Santa Clara, CA, USA). A new immunohistochemical algorithm was then developped and validated in concordance with GEP using a training set and a validation set (100 cases for training, 50 cases for validation). The first step was to choose the best and most parsimonious machine learning algorithm capable of predicting DLBCL subtypes. Then, logical rules were created based on the predicted probabilities of class membership using a probability threshold that optimize concordance between machine learning and logical rules predictions. xgBoost algorithm (boosted trees models) with 4 covariables was chosen, using a cut-off of 0.39 for predicted class membership probabilities.ResultsA new algorithm (Table 1) using CD10, MUM1, FOXP1 and IGM was then derived with an accuracy of 93% (IC95 [0.80 ; 0.98]) on the validation set, compared to GEP classification. This algorithm has four arms taking into account the GCB marker CD10 and independently one of the three ABC (MUM1, FOXP1, IGM) markers to classify as GCB, ABC or non classified. Hans algorithm had an accuracy of 0.88 (IC95 [0.75 ; 0.96]) on this same dataset. On the training and validation set combined, our algorithm re-classified correctly 7 GCB subtypes compared to the Hans algorithm.Survival analysis on the remaining cases (611 patients) found an ABC vs GC hazard ratio (HR) of 2.0 (IC95 [1.47 ; 2.73]) for progression free survival (PFS) and 1.95 (IC95 [1.36-2.80]) for overall survival (OS), using ABC vs GCB classification of our new algorithm. In our series, Hans' algorithm found an ABC vs GCB HR of 1,74 IC95 [1,27-2,39] for PFS and 1,87 IC95 [1,29-2,71] for OS.ConclusionOur new algorithm shows excellent performance compared to the Gold Standard GEP classification, using only 4 immunostains, and allows for more accurate risk stratification both for PFS and OS than Hans'algorithm. Therefore, it can be used as a useful tool in addition to molecular signatures. [Display omitted] DisclosuresBruneau:Innate Pharma: Research Funding. Delarue:Gilead: Consultancy, Honoraria; BMS: Consultancy, Honoraria; Roche: Consultancy, Honoraria. Haioun:PFIZER: Consultancy, Honoraria; GILEAD: Consultancy, Honoraria; JANSSEN: Consultancy, Honoraria; Sandoz: Consultancy, Honoraria; Takeda: Consultancy, Honoraria; Amgen: Consultancy, Honoraria; Roche: Consultancy, Honoraria. Casasnovas:BMS: Consultancy, Honoraria; Abbvie: Consultancy, Honoraria; Sanofi: Consultancy, Honoraria; Takeda: Consultancy, Honoraria; Gilead: Consultancy, Honoraria, Research Funding; Roche: Consultancy, Honoraria, Research Funding. Morschhauser:Roche: Consultancy, Honoraria; Celgene: Consultancy, Honoraria; Janssen: Consultancy, Honoraria; Bristol-Myers Squibb: Consultancy, Honoraria; Gilead: Consultancy; Servier: Consultancy. Emile:Merck Serono: Honoraria. Tilly:Gilead: Honoraria; Immunogen: Honoraria; Karyopharm: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees; Celgene: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; Roche: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; Janssen: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees; Takeda: Consultancy, Honoraria. Salles:Celgene: Consultancy, Honoraria; Gilead: Consultancy, Honoraria; Janssen: Consultancy, Honoraria; Kite: Consultancy, Honoraria; MSD: Consultancy, Honoraria; Novartis: Consultancy, Honoraria; Servier: Consultancy, Honoraria; BMS: Consultancy; morphosys: Consultancy, Honoraria; Amgen: Consultancy, Honoraria; Roche: Consultancy, Honoraria, Research Funding. Jardin:Roche: Honoraria; celgen: Honoraria; Janssen: Honoraria. Molina:Merck Serono: Honoraria; Novartis: Honoraria; Takeda: Other: travel support to the ASH meeting.

Utilizing Digital Pathology Informatics Algorithms for Diffuse Large B-Cell Lymphoma Subtyping

Introduction:Diffuse large B-cell lymphoma (DLBCL) is the most commonly occurring lymphoid malignancy in the United States. Through gene expression profiling (GEP), two biologically distinct molecular subgroups of DLBCL have been identified - germinal center B-cell (GCB) and activated B-cell - with disparate prognostic implications. Because GEP is not yet easily implemented in clinical practice, immunohistochemical (IHC) algorithms have been developed and validated to segment DLBCL into cell-of-origin subtypes. The most commonly employed algorithm uses CD10, BCL6 and MUM1 to distinguish GCB and non-GCB subtypes, but has been shown to misclassify up to 20% of cases. Advancements in digital pathology technology have led to the development of image analysis algorithms that allow for the extraction of quantitative descriptions of pathology image features such as level of IHC staining. We explore the feasibility of developing a precise, repeatable, and objective image analysis-based scoring algorithm for DLBCL whole slide images to replicate DLBCL subtype classification and prognostic assessment.Methods:Tissue slides of the immunostains included in the Hans algorithm (CD10, BCL6, MUM1) for 40 DLBCL patients were digitized at 40x objective resolution using Hamamatsu Nanozoomer 2.0 HT or Aperio AT2. We developed and trained an image analysis algorithm on the IHC whole slide images. The algorithm first masks the tissue area of the whole slide image, then performs superpixel segmentation, dividing tissue area into compact regions (60 microns square). Each region is normalized to a reference IHC image and has a color deconvolution algorithm applied to separate the hematoxylin and diaminobenzidine (DAB) stains. Features describing the distribution of both the hematoxylin and DAB stains are extracted using convolutional autoencoders. We then trained a classification algorithm using these features to classify the staining of these regions as positive or negative using machine learning techniques. Whereas pathologists examine specific regions in a high-power field and use visual estimation to determine if >30% of tumor cells are positive, our algorithm measures the percentage of positive regions over the entire DLBCL tissue. Receiver operating characteristics (ROC) curves were calculated to assess ability of image analysis-based measurements to predict pathologist classifications as positive or negative for CD10, BCL6, and MUM1. Pathologist classification of DLBCL subtype and positive or negative for each immunostain were obtained from pathology reports. Using thresholds for percent positive regions established by the ROC curves, patients were classified into GCB/non-GCB by the Hans algorithm using sequential application of the identified thresholds as shown in the Figure. The regions in the figure define the algorithm classification and the lines correspond to the pathologist classification displaying GCB in blue and non-GCB in red. Concordance with pathologist subtyping was calculated by agreement rate and Cohen's κ statistic.Results:Of the 40 patients included in our study, pathologists classified CD10, BCL6, and MUM1 as positive in 27.5%, 90%, and 84.9% respectively. Area under the ROC curve (AUC) for predicting pathologist classifications of positivity from image analysis measurements for CD10, BCL6, and MUM1 were 0.92, 1.0, and 0.95 respectively. Thresholds from the ROC curves for CD10, BCL6, and MUM1 were 13%, 15%, and 20% respectively. Using these thresholds, classification by image analysis algorithm was concordant with pathologist classification in 82.5% (κ = 0.65) of cases.Discussion:Pathologist IHC classification is limited by interobserver and interlaboratory variability, human error, difficulties with human eye distinguishing between low-intensity stains, and ordinal outputs. These limitations have led to inconsistent prognostic significance with IHC cell-of-origin subtyping. We developed a novel, objective image analysis scoring for predicting DLBCL subtype that uses whole slide images. Our image analysis algorithm may provide an effective decision-support tool for pathologist subtype classification and address known limitations. Future work will compare subtyping by the image analysis algorithm to gene expression profiling and inter-rater reliability of tissue microarrays and whole slide images performed by expert hematopathologists. [Display omitted] DisclosuresFlowers:National Institutes Of Health: Research Funding; Burroughs Welcome Fund: Research Funding; Celgene: Consultancy, Research Funding; Gilead: Consultancy; Research to Practice: Research Funding; Infinity: Research Funding; Onyx: Research Funding; Seattle Genetics: Consultancy; V Foundation: Research Funding; Genentech/Roche: Consultancy, Research Funding; OptumRx: Consultancy; Bayer: Consultancy; Millennium/Takeda: Research Funding; Eastern Cooperative Oncology Group: Research Funding; Clinical Care Options: Research Funding; National Cancer Institute: Research Funding; TG Therapeutics: Research Funding; Acerta: Research Funding; Prime Oncology: Research Funding; Educational Concepts: Research Funding; Spectrum: Consultancy; Pharmacyclics LLC, an AbbVie Company: Research Funding; Janssen Pharmaceutical: Research Funding; Abbvie: Consultancy, Research Funding. Cooper:Varian Medical Systems: Honoraria.

Tumor infiltrating lymphocytes (TILs) in endometrioid and clear cell ovarian carcinoma: Characterization, association with mismatch repair system deficiency and endometriosis, and prognostic implications.

e17549 Background: In contrast to high-grade serous ovarian cancer, relationship between tumor infiltrating lymphocytes (TILs) with endometrioid (EC) and clear cell ovarian carcinomas (CC) are less known. However, these tumors are thought to arise from a transformed endometriosis, a disease with a marked inflammatory response, and are also more related to Lynch syndrome. The characterization of TILs in these tumors could lead to stratify patients according to their prognosis and to identify candidates for immumotherapy in case of relapse. Methods: We conducted a retrospective observational study of 121 patients with EC and CC, diagnosed and treated in La Paz University Hospital, from November 1992 until December 2013. An expert pathologist reviewed the histopathological diagnosis, using the immunohistochemical algorithm with WT1, p53, Napsin A and progesterone receptor, and the presence of endometriosis. Clinicopathological data, genetic testing and personal and familial background were collected. Tissue-microarrays (TMAs) were used to analyze CD8+, CD4+ and CD3+ intra-epithelial TILs, according to the score designed by the Ovarian Tumor Tissue Analysis Consortium, peritumoral and endometriosis lymphocytes, ARID1A, MLH1, PMS2, MSH2 and MSH6. Software SPSS18 version was used for descriptive and statistical analysis. Results: Mean age was 54.3 years, and the median follow-up 10.02 years. 48.8% were CC and 51.2% EC, of which, 55.9% were well-differentiated, 32.2% moderately and 11.9% poorly. 58.8% of patients were diagnosed in early stages and 41.2% in advanced ones. 9.1% had family history of ovarian, colorectal or endometrial cancer and 16.5% of breast cancer. 37.2% of patients had a relapse (median progression-free survival of 31 months) and 41.6% died (32.6% related to the tumor). 55.4% of the patients had endometriosis (53.2% were EC and 57.6% CC). 8.2% had lost of MLH1, PMS2, MSH2 or MSH6 and 42% lost ARID1A expression. Moderate or high levels of CD3+, CD8+ and CD4+ TILs were seen in the 57.9%, 53.7% and 1.1%, respectively. There was no relationship between TILs and the lost of MLH1, PMS2, MSH2 and MSH6. Moderate or higher levels of CD3+ TILs were related with better overall survival (OS) only in early stages (p = 0.045). Increased CD8+ TILs in EC patients were associated with a trend towards an improved OS in comparison with those patients with lower or negative CD8+ TILs. Moderate o higher levels of CD8+ TILs were also associated, without statistical significance, with the presence of endometriosis and ARID1A. Conclusions: In our study CD3+ TILs correlated to a better OS, but only in early stages of CC and EC. CD8+ TILs were also associated with a tendency to better OS in EC. We did not find statistically significant association between TILs and Mismatch Repair System deficiency, ARID1A or endometriosis.

Simultaneous Identification of Cell of Origin, Translocations, and Hotspot Mutations in Diffuse Large B-Cell Lymphoma Using a Single RNA-Sequencing Assay.

Diffuse large B-cell lymphoma (DLBCL) is an aggressive non-Hodgkin lymphoma with a heterogenous genetic landscape that can require multiple assays to characterize. We reviewed a 1-step RNA-based assay to determine cell of origin (COO), detect translocations, and identify mutations and to assess the role of the assay in diagnosis. Using a single custom Archer FusionPlex Lymphoma panel, we performed anchored multiplex polymerase chain reaction-based RNA sequencing on 41 cases of de novo DLBCL. Each case was subclassified by COO, and gene fusions and hotspot mutations were identified. The findings were then compared with COO classification by the Hans immunohistochemical algorithm and NanoString technology, cytogenetics, and fluorescence in situ hybridization results. Concordant COO classification by the FusionPlex panel and NanoString was observed in 35 of 41 cases (85.3%), with NanoString and Hans concordant in 33 of 41 cases (80.5%) and FusionPlex and Hans concordant in 33 of 41 cases (80.5%). The FusionPlex assay also detected 6 of 11 BCL6 translocations (4 cryptic), 2 of 3 BCL2 translocations, and 2 of 4 MYC translocations. Mutations were detected in lymphoma-related genes in 24 of 41 cases. This FusionPlex assay offers a single method for COO classification, mutation detection, and identification of important translocations in DLBCL. Although not replacing traditional testing, it could offer useful data when limited tissue is available.

Exosomes and extracellular vesicles as liquid biopsy biomarkers in diffuse large B-cell lymphoma: Current state of the art and unmet clinical needs.

Diffuse large B-cell lymphoma (DLBCL) is the most common type of non-Hodgkin's lymphoma, and it constitutes biologically heterogeneous entities. Standard first-line therapies cure ~60% of patients, the rest being either refractory or experiencing relapse. Currently, there are no robust predictive biomarkers of therapeutic response. Heterogeneity of DLBCL is partly explained by the cell of origin (COO), ie, germinal centre B cell or activated B cell, with the latter exhibiting worse prognosis. While gene expression profiling (GEP) is the gold standard for determining COO, surrogate immunohistochemical algorithms are used clinically, but show significant discordance with GEP. Recently, additional genetic subgroups with different prognoses have been reported. However, the tools/expertise required for analysis prohibit widespread deployment. Liquid biopsy-based assays show promise in providing clinically actionable information, are noninvasive and facilitate serial sampling to assess mechanisms of therapy resistance. Circulating, cell-free DNA analysis has shown enhanced sensitivity for detecting molecular alterations, but this modality cannot determine alterations of the tumor proteome or on signalling pathways. Exosomes are endosomally derived vesicles, are found in high abundance in body fluids and are readily isolated using a variety of methods. Tumour-derived exosomes can yield data regarding genetic, transcriptional, and proteomic changes useful for diagnosis, prognosis, and therapy of DLBCL. At present, standardized techniques for isolating exosomes are lacking and discriminating between exosomes from neoplastic and normal B cells is challenging. Refinements in isolation procedures are required to realize their full potential as precision medicine tools to provide comprehensive information on disease subtypes, identify prognostic factors, allow real-time monitoring of therapy response and delineate novel drug targets.

Performance of the Lymph2Cx Cell of Origin Classifier of Diffuse Large B-Cell Lymphoma in Comparison to Two Immunohistochemical Algorithms

Performance of the Lymph2Cx Cell of Origin Classifier of Diffuse Large B-Cell Lymphoma in Comparison to Two Immunohistochemical Algorithms

Abstract B10: A simple immunohistochemical algorithm to evaluate molecular subtypes and clinical associations in non-muscle invasive bladder cancer

Abstract Introduction: Recent large-scale transcriptomic studies in muscle-invasive bladder cancer (MIBC) have identified promising predictive and prognostic properties of luminal, basal, and additional molecular subtypes. Transcriptomic profiling is laborious, expensive, and prone to errors due to infiltrating lymphocytes and variations in tumor cellularity. Immunohistochemistry (IHC) provides a faster, cheaper, and clinically feasible subtyping method that has been validated to subtype MIBC (1) and simplified into a three-antibody algorithm (2). Here we utilize this algorithm in non-muscle invasive bladder cancer (NMIBC), where, in contrast to MIBC, there are few data exploring IHC subtypes. Methods: We implemented the simplified three-antibody IHC-based algorithm to subtype a cohort of 523 non-muscle invasive (stages pTa low grade, pTa high grade, and pT1 high grade) transurethral resections of bladder tumor (TURBT) samples. Cases were sampled in duplicate on six tissue microarray blocks. Spatial localization, prevalence, and intensity of antibody staining was scored using digital image analysis. Results: Similar to previous transcriptome-based work (3), preliminary analysis identified the prevalence of Urothelial-Like (Uro-like), Genomically Unstable (GU), Basal-Squamous Cell Carcinoma-Like (BSq), and unclassified subtypes at 33% n=171, 48%, n=251, 9%, n=46, and 10%, n= 55, respectively. As expected, GU was enriched in pTa high grade and pT1 tumors, but surprisingly, BSq cancers were more similar to Uro-like in stage and grade distribution. Discussion: Molecular subtypes of bladder cancer can be identified easily using a three-antibody assay. Subtypes that signify high risk of metastasis in MIBC are surprisingly prevalent in NMIBC but may not yet have acquired the molecular lesions that drive metastatic progression. Further research into subtype and progression should be accelerated by the availability of rapid and inexpensive IHC-based assays.

Immunohistochemical Study of Dark and Progenitor Carotid Body Cells: Artefacts or Real Subtypes?

Postmortem changes occurring in human carotid body were simulated on the Wistar rat model. It was shown that light, dark, and pyknotic (progenitor) subtypes of human carotid body cells are an artifact and cannot be used in clinical practice to study the characteristics of various human diseases. The differences between the control group of healthy individuals and individuals with the various pathologies are most likely due to the different levels of premortal hypoxia that the tissue had been exposed to. Moreover, widespread antigens used in practice were divided into 2 groups by their tolerance to autolysis: stable and unstable ones. This can be useful for the development of immunohistochemical test algorithms for the diagnostics on autopsy material.

Molecular Subtypes and Genomic Profile of Primary Central Nervous System Lymphoma.

Primary central nervous system lymphomas (PCNSL) are aggressive non-Hodgkin lymphomas affecting the central nervous system (CNS). Although immunophenotyping studies suggested an uniform activated B-cell (ABC) origin, more recently a spectrum of ABC and germinal center B-cell (GC) cases has been proposed, with the molecular subtypes of PCNSL still being a matter of debate. With the emergence of novel therapies demonstrating different efficacy between the ABC and GC patient groups, precise assignment of molecular subtype is becoming indispensable. To determine the molecular subtype of 77 PCNSL and 17 secondary CNS lymphoma patients, we used the NanoString Lymphoma Subtyping Test (LST), a gene expression-based assay representing a more accurate techniqueof subtyping compared with standard immunohistochemical (IHC) algorithms. Mutational landscapes of 14 target genes were determined using ultra-deep next-generation sequencing. Using the LST-assay, a significantly lower proportion (80% vs 95%) of PCNSL cases displayed ABC phenotype compared with the IHC-based characterization. The most frequently mutated genes included MYD88, PIM1, and KMT2D. In summary, we successfully applied the LST-assay for molecular classification of PCNSL, reporting higher proportion of cases with GC phenotype compared with IHC analyses, leading to a more precise patient stratification potentially applicable in the diagnostic algorithm of PCNSL.