Modern Pathology最新文献

{"title":"Statistics of Generative AI & Non-Generative Predictive Analytics Machine Learning in Medicine.","authors":"Hooman H Rashidi, Bo Hu, Joshua Pantanowitz, Nam Tran, Silvia Liu, Alireza Chamanzar, Mert Gur, Chung-Chou H Chang, Yanshan Wang, Ahmad Tafti, Liron Pantanowitz, Matthew G Hanna","doi":"10.1016/j.modpat.2024.100663","DOIUrl":"https://doi.org/10.1016/j.modpat.2024.100663","url":null,"abstract":"<p><p>The rapidly evolving landscape of artificial intelligence (AI) and machine learning (ML) in medicine has prompted medical professionals to increasingly familiarize themselves with related topics. This also demands grasping the underlying statistical principles that govern their design, validation, and reproducibility. Uniquely, the practice of pathology and medicine produces vast amount of data that can be exploited by AI/ML. The emergence of generative AI, especially in the area of large language models and multimodal frameworks, represents approaches that are starting to transform medicine. Fundamentally, generative and traditional (e.g., non-generative predictive analytics) ML techniques rely on certain common statistical measures to function. However, unique to generative AI are metrics such as, but not limited to, perplexity and BiLingual Evaluation Understudy (BLEU) score that provide a means to determine the quality of generated samples that are typically unfamiliar to most medical practitioners. In contrast, non-generative predictive analytics ML often employs more familiar metrics tailored to specific tasks as seen in the typical classification (i.e., confusion metrics measures such as accuracy, sensitivity, F1-score, ROC-AUC, etc.) or regression studies (i.e., Root mean Square Error [RMSE], R-squared, etc.). To this end, the goal of this review article (as part 4 of our AI review series) is to provide an overview and comparative measure of statistical measures and methodologies employed in both generative AI and traditional (i.e., non-generative predictive analytics) ML fields, along with their strengths and known limitations. By understanding their similarities and differences along with their respective applications, we will become better stewards of this transformative space which ultimately enables us to better address our current and future needs and challenges in a more responsible and scientifically sound manner.</p>","PeriodicalId":18706,"journal":{"name":"Modern Pathology","volume":" ","pages":"100663"},"PeriodicalIF":7.1,"publicationDate":"2024-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142695549","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"RB Loss in p53 Abnormal Endometrial Carcinoma: Histological and Clinicopathological Correlates.","authors":"Ezgi Dicle Serbes, Nanda Horeweg, Carlos Parra-Herran, Renske van Rijnsoever, Jan J Jobsen, Ina Jurgenliemk-Schulz, Nienke Kuijsters, Remi A Nout, Marie A D Haverkort, Melanie E Powell, Pearly Khaw, Marie Plante, Catherine Genestie, Hans W Nijman, Carien L Creutzberg, Tjalling Bosse, Claire J H Kramer","doi":"10.1016/j.modpat.2024.100660","DOIUrl":"https://doi.org/10.1016/j.modpat.2024.100660","url":null,"abstract":"<p><p>Of the four molecular subtypes of endometrial cancer (EC), p53-abnormal (p53abn) EC is associated with abundant copy number alterations (CNAs) and the worst clinical outcome. Patients with p53abn EC have the highest risk of disease recurrence and death, independent of tumor grade and histologic subtype. Currently, all invasive p53abn ECs are considered high risk, and no prognostic biomarkers have yet been found that can aid in clinical management. Here, we aimed to test whether loss of retinoblastoma protein (RB) expression using immunohistochemistry (IHC) has potential for prognostic refinement of p53abn EC. A large cohort of 227 p53abn ECs collected from the PORTEC-1/2/3 clinical trials and the MST cohort study was investigated, and RB loss was identified in 7.0% (n=16/227). RB-lost p53abn ECs were predominantly high-grade endometrioid ECs (n=6, 37.5%) and carcinosarcomas with endometrioid-type epithelial component (n=5, 31.3%). Histologically, RB-lost p53abn EC were typified by high grade nuclear atypia (n=16, 100%), predominantly solid growth pattern (n=15/16, 93.8%), and polypoid growth (n=9/16, 56.3%). Copy number loss involving the RB1 locus was identified in the majority of RB-lost p53abn EC (n=13/14, 92.9%), explaining the loss of RB expression. Comparative analysis also showed that RB-lost p53abn EC were diagnosed at earlier stages than RB-retained p53abn EC (p=0.014). Interestingly, RB-lost p53abn EC showed prolonged time to overall recurrence (p=0.038), even within stage I alone (p=0.040). These findings highlight distinct morphomolecular features in RB-lost p53abn EC and confirm the utility of RB IHC as a surrogate for molecular RB1 alterations. This is the first study to show the potential use of RB in prognostic refinement of p53abn EC, although validation is warranted.</p>","PeriodicalId":18706,"journal":{"name":"Modern Pathology","volume":" ","pages":"100660"},"PeriodicalIF":7.1,"publicationDate":"2024-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142693201","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Extra-Axial Poorly Differentiated Chordoma: Clinicopathologic and Molecular Genetic Characterization.","authors":"William J Sande, Andrew L Folpe, Paige O'Connor, Daniel Graham, Jeremy F Molligan, Ying-Chun Lo, Yvonne Y Cheung, Baptiste Ameline, Daniel Baumhoer, Dorothee Harder, Kevin A Raskin, Christopher W Mount, Yin P Hung, G Petur Nielsen, Darcy A Kerr, Darya Buehler, Doris E Wenger, Judith Jebastin Thangaiah","doi":"10.1016/j.modpat.2024.100664","DOIUrl":"https://doi.org/10.1016/j.modpat.2024.100664","url":null,"abstract":"<p><p>Poorly differentiated chordoma (PDC) is an aggressive subtype of chordoma characterized by SMARCB1 (INI1) loss and a dismal prognosis. It typically involves the axial skeleton, most commonly the skull base and the cervical spine. To our knowledge, only five cases of extra-axial PDC (EAPDC) have been reported, and the natural history of these tumors is not fully understood. We studied six cases of extra-axial poorly differentiated chordoma, with the goal of better understanding these exceptionally rare tumors. The tumors occurred in four women and two men, ranging from 37 to 68 years of age (median 57.5 years) and involved or originated in the left knee joint (3 cases) and right knee joint (2 cases) and right wrist (1 case). Grossly, all were solid and lobulated, with areas of necrosis. Histologically, the tumors were identical to axial poorly differentiated chordoma, with sheets and lobules of overtly malignant-appearing epithelioid-to-rhabdoid cells with prominent nucleoli. Mitotic activity and necrosis were present. By immunohistochemistry, all cases expressed keratins and brachyury and were SMARCB1-deficient. Molecular genetic analysis identified SMARCB1 loss-of-function alterations in 4 of the tested cases, including mutations (2 cases) and copy number loss (2 cases). DNA methylation profiling of 4 cases of extra-axial poorly differentiated chordoma showed clustering with axial PDC. Clinical follow-up (6 patients; median 11.5 months; range 1-26 months) showed 4 patients to have received transfemoral amputation and one extra-articular resection. None received neoadjuvant radiotherapy; one received neoadjuvant chemotherapy, one adjuvant chemotherapy/immunotherapy. Local recurrences were seen in 2 patients at 7 and 8 months; 3 patients developed metastases 7-11 months after surgery. Two patients were alive with metastatic disease (at 7 and 13 months), one dead of disease (20 months) and three disease-free (1- 26 months). We conclude that extra-axial poorly differentiated chordoma are aggressive malignancies with an unusual predilection for the knee joint and unknown pathogenesis.</p>","PeriodicalId":18706,"journal":{"name":"Modern Pathology","volume":" ","pages":"100664"},"PeriodicalIF":7.1,"publicationDate":"2024-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142693198","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Fast Processing of Electron Microscopic Specimen Preserved Ultrastructure of Glomeruli and Electron Dense Deposits in Diagnostic Renal Biopsies: A Prospective and Retrospective Comparative Study.","authors":"Aranza Pinedo, Prerna Rastogi, Abdullah Thayyil, Matthew Gosse, Amy Trent, Serena M Bagnasco, Avi Rosenberg, Lois J Arend, Dao-Fu Dai","doi":"10.1016/j.modpat.2024.100662","DOIUrl":"https://doi.org/10.1016/j.modpat.2024.100662","url":null,"abstract":"<p><p>Optimization of electron microscopy (EM) tissue processing protocols is essential to handle the global increase in the number of renal biopsies requiring EM for accurate diagnosis. The conventional EM processing method (CEM) is the standard method used by >95% of laboratories worldwide and it takes at least 48-52 hours for completion. In contrast, a fast-processing EM method (FEM) using microwave irradiation can be completed in 8 hours, allowing EM findings to be reported within 24 hours for most cases. There is widespread concern about the suboptimal quality of the FEM that may compromise its diagnostic roles; however, qualitative and quantitative data supporting the non-inferiority of FEM compared to CEM has not been reported. We performed both prospective and retrospective studies. The prospective analysis compares FEM and CEM images from the same biopsy samples. For each case, the tissue was divided into two pieces; one piece for FEM processing and the second for CEM processing. The retrospective study compares the EM images of renal cases with electron-dense deposits from our archives that were processed either by FEM or CEM. The prospective analysis included 4 cases: lupus membranous nephropathy, IgA nephropathy, immune-complex mediated glomerulonephritis, and acute tubular injury. Both FEM and CEM methods obtained high-resolution images with comparable quality. A quantitative morphometric analysis of the glomerular basement membrane (GBM) in the IgA nephropathy case showed similar GBM thickness when processed by the FEM and the CEM, suggesting that FEM did not affect GBM thickness. The retrospective study of 42 cases with electron-dense deposits showed that the ultrastructural features of electron-dense deposits were indistinguishable between the FEM and the CEM. This included microtubular substructures in immunotactoid glomerulonephritis, the \"fingerprint\" deposits in cryoglobulinemic glomerulonephritis, fibril deposits in the light-chain amyloidosis as well as fibrillary glomerulonephritis, with comparable morphometric measurements of the deposits. The FEM is efficient, consistent, reproducible, and delivers comparable high-quality sections and images for diagnostic assessment of renal biopsies, comparable to those attained by the CEM while decreasing turnaround time significantly, making it possible to provide faster and accurate diagnostic results.</p>","PeriodicalId":18706,"journal":{"name":"Modern Pathology","volume":" ","pages":"100662"},"PeriodicalIF":7.1,"publicationDate":"2024-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142693199","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Congenital peribronchial myofibroblastic tumors harbor a recurrent EGFR kinase domain duplication.","authors":"Sheren Younes, Carlos J Suarez, Jennifer Pogoriler, Tricia Bhatti, Archana Shenoy, Raya Saab, Lea F Surrey, Serena Y Tan","doi":"10.1016/j.modpat.2024.100661","DOIUrl":"https://doi.org/10.1016/j.modpat.2024.100661","url":null,"abstract":"<p><p>Congenital peribronchial myofibroblastic tumor (CPMT) is a rare benign infantile pulmonary neoplasm that presents prenatally or early in infancy, and exhibits distinctive histologic features characterized by the presence of cartilaginous islands intermixed with bland spindle cells, not uncommonly displaying prominent mitoses. Despite its benign nature, CPMT can lead to fetal demise, postnatal respiratory distress, or complications from perinatal surgical resection. Although the morphologic and clinical features of CPMT are well-described, its molecular features and oncogenesis remain elusive. Following the detection of EGFR kinase domain duplication (KDD) of exons 18-25 in an index case, we identified three additional cases of morphologically classic and clinically well-characterized CPMTs from the archives and performed targeted RNA- and DNA-based profiling via next generation sequencing for detection of rearrangements, sequence variants and copy number variants on all cases. Two cases were detected prenatally, one patient presented at birth and one at 8 weeks of life. All tumors were resected, with follow-up period ranging from 0 days to 10 years. One patient died shortly after surgical resection, and the other three have had no recurrences. In all cases, EGFR KDD was detected. In two of four cases, gains of select whole chromosomes were noted. Our findings establish EGFR KDD as a recurrent oncogenic driver of CPMT. Notably, this alteration is also found in classical congenital mesoblastic nephromas (CMNs), infantile kidney tumors with which CPMTs share striking morphologic and clinical similarities. This strongly suggests CPMTs and classical CMNs share common oncogenesis, and represent the same tumor in different locations. EGFR KDDs have also been reported in neonatal soft tissue tumors with infantile fibrosarcoma-like histology and cartilaginous differentiation, raising questions about their relationship. EGFR KDD emerges as a diagnostic marker, potential therapeutic target, and a window into the oncogenesis of a distinct subset of infantile mesenchymal tumors.</p>","PeriodicalId":18706,"journal":{"name":"Modern Pathology","volume":" ","pages":"100661"},"PeriodicalIF":7.1,"publicationDate":"2024-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142693123","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Spread Through Air Spaces: Comment","authors":"Hinpetch Daungsupawong ,&nbsp;Viroj Wiwanitkit","doi":"10.1016/j.modpat.2024.100634","DOIUrl":"10.1016/j.modpat.2024.100634","url":null,"abstract":"","PeriodicalId":18706,"journal":{"name":"Modern Pathology","volume":"37 12","pages":"Article 100634"},"PeriodicalIF":7.1,"publicationDate":"2024-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142639271","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Making pathologists ready for the new AI era: changes in required competencies.","authors":"Shoko Vos, Konnie Hebeda, Megan Milota, Martin Sand, Jojanneke Drogt, Katrien Grünberg, Karin Jongsma","doi":"10.1016/j.modpat.2024.100657","DOIUrl":"10.1016/j.modpat.2024.100657","url":null,"abstract":"<p><p>In recent years, there has been an increasing interest in developing and using artificial intelligence (AI) models in pathology. While pathologists generally have a positive attitude towards AI, they report a lack of knowledge and skills regarding how to use it in practice. Furthermore, it remains unclear what skills pathologists would require to use AI adequately and responsibly. Yet adequate training of (future) pathologists is essential for successful AI use in pathology. In this paper, we assess which entrustable professional activities (EPAs) and associated competencies pathologists should acquire in order to use AI in their daily practice. We make use of available academic literature, including literature in radiology, another image-based discipline, which is currently more advanced in terms of AI development and implementation. Although microscopy evaluation and reporting could be transferrable to AI in the future, most of the current pathologist EPAs and competencies will likely remain relevant when using AI techniques and interpreting and communicating results for individual patient cases. In addition, new competencies related to technology evaluation and implementation will likely be necessary, and knowing one's own strengths and limitations in human-AI interaction. Because current EPAs do not sufficiently address the need to train pathologists in developing expertise related to technology evaluation and implementation, we propose a new EPA to enable pathology training programs to make pathologists fit for the new AI era \"using AI in diagnostic pathology practice\" and outline its associated competencies.</p>","PeriodicalId":18706,"journal":{"name":"Modern Pathology","volume":" ","pages":"100657"},"PeriodicalIF":7.1,"publicationDate":"2024-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142623644","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Primary vulvar and vaginal adenocarcinomas of intestinal-type are closer to colorectal adenocarcinomas than to carcinomas of Müllerian origin.","authors":"Alexis Trecourt, Isabelle Treilleux, Daniel Pissaloux, Marie Donzel, Brice Thamphya, Franck Thirode, Aurélie Houlier, Sandrine Paindavoine, Tatiana Franceschi, Aline Baltrès, Witold Gertych, Pierre-Adrien Bolze, Pierre Antoine Chaix, Charlotte Roux-Terrier, Françoise Descotes, Isabelle Ray-Coquard, Jonathan Lopez, Mojgan Devouassoux-Shisheboran","doi":"10.1016/j.modpat.2024.100649","DOIUrl":"https://doi.org/10.1016/j.modpat.2024.100649","url":null,"abstract":"<p><p>Primary vulvar and vaginal adenocarcinomas of intestinal-type (VVAIts) are very rare tumors, displaying morphological and immunohistochemical overlap with colorectal adenocarcinomas. However, their immunoprofile and genomics are poorly studied, and their origin is still debated. Herein, we studied a series of 4 vulvar and 4 vaginal adenocarcinomas of intestinal-type using a large panel of immunohistochemistry, and DNA and RNA sequencing with clustering analyses. All tumors shared a similar morphology with colorectal adenocarcinomas and diffuse CK20 and CDX2 expression. SATB2 diffuse positivity was observed in 62.5% of tumors and CK7 in 82.5%, while PAX8, SOX17, p16, estrogen and progesterone receptors were always negative. A p53 mutated-type expression was observed in 75% of tumors. All tumors were MMR proficient. No HPV DNA nor pathogenic transcript fusions were detected. The most frequent molecular alterations were TP53 and KRAS variants in 71.4% and 42.9%, respectively. The transcriptomic analysis highlighted a robust VVAIts cluster distinct from endocervical, ovarian, lung, thyroid, salivary glands, breast, and renal carcinomas, but failed to differentiate vulvar from vaginal intestinal-type tumors. On two different clustering analyses, VVAIts clustered altogether, very close to colorectal adenocarcinomas. Compared to endocervical adenocarcinomas of intestinal-type, VVAIts had a significantly lower expression of SOX17 and epithelial-mesenchymal transition (EMT) genes, and a higher MAP-Kinase pathway gene expression. These results suggest that Müllerian structures leading to cervical adenocarcinomas may undergo intestinal-type transdifferentiation via an EMT phenomenon. Conversely, MAP-Kinase pathway activation in VVAIts, which plays a major role in colorectal adenocarcinomas, may indicate a close relationship in the carcinogenesis of these tumors. Our results indicate that adenocarcinomas of intestinal-type, in distal vagina or vestibular vulva, might be a unique and single entity, probably originating from cloacogenic embryonic remnants and/or ectopic colorectal mucosae inclusions. An open question would be to explore the efficacy of systemic drugs prescribed in colorectal cancers, in VVAIts.</p>","PeriodicalId":18706,"journal":{"name":"Modern Pathology","volume":" ","pages":"100649"},"PeriodicalIF":7.1,"publicationDate":"2024-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142623664","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Weakly-supervised classification of Mohs surgical sections using artificial intelligence.","authors":"Daan J Geijs, Lisa M Hillen, Stephan Dooper, Véeronique Winnepenninckx, Vamsi Varra, David R Carr, Kathryn T Shahwan, Geert Litjens, Avital Amir","doi":"10.1016/j.modpat.2024.100653","DOIUrl":"https://doi.org/10.1016/j.modpat.2024.100653","url":null,"abstract":"<p><p>Basal cell carcinoma (BCC) is the most frequently diagnosed form of skin cancer, and its incidence continues to rise, particularly among older individuals. This trend puts a significant strain on healthcare systems, especially in terms of histopathologic diagnostics required for Mohs micrographic surgery (MMS), which is used to treat BCC in sensitive locations to minimize tissue loss. This study aims to address the challenges in BCC detection within MMS whole-slide images (WSIs) by developing and evaluating a deep learning model that bridges weakly-supervised learning with interpretable segmentation-based methods through attention maps. Utilizing datasets from two medical centers, the model demonstrated an average area under the ROC curve (AUC) of 0.958 on internal testing and an AUC of 0.934 on an independent third external dataset despite no fine-tuning or preprocessing for the latter. Attention maps provided insights into the model's decision-making, highlighting critical regions for slide-level classification. The sensitivity of the attention maps in localizing tumor regions was 0.853 when no filtering was applied and gave 8 revision false positives per slide on average and was reduced to an average of 2 false positives per slide with a sensitivity of 0.873 when detections smaller than 200 micrometers were removed from the attention maps. These findings indicate that the deep learning model is highly effective in detecting BCC in MMS WSIs, with robust performance across different datasets and conditions. The use of attention maps enhances the model's interpretability, making it a promising tool for aiding dermatopathologists and MMS surgeons.</p>","PeriodicalId":18706,"journal":{"name":"Modern Pathology","volume":" ","pages":"100653"},"PeriodicalIF":7.1,"publicationDate":"2024-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142623914","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Spatial Architecture of Single-cell and Vasculature in Tumor Microenvironment Predicts Clinical Outcomes in Triple-Negative Breast Cancer.","authors":"Haoyang Mi, Ravi Varadhan, Ashley M Cimino-Mathews, Leisha A Emens, Cesar A Santa-Maria, Aleksander S Popel","doi":"10.1016/j.modpat.2024.100652","DOIUrl":"10.1016/j.modpat.2024.100652","url":null,"abstract":"<p><p>Triple-negative breast cancer (TNBC) is an aggressive subtype of breast cancer with limited treatment options, which warrants the identification of novel therapeutic targets. Deciphering nuances in the tumor microenvironment (TME) may unveil insightful links between anti-tumor immunity and clinical outcomes, yet such connections remain underexplored. Here we employed a dataset derived from imaging mass cytometry of 71 TNBC patient specimens at single-cell resolution and performed in-depth quantifications with a suite of multi-scale computational algorithms. The TNBC TME reflected a heterogeneous ecosystem with high spatial and compositional heterogeneity. Spatial analysis identified ten recurrent cellular neighborhoods (CNs) - a collection of local TME characteristics with unique cell components. The prevalence of CNs enriched with B cells, fibroblasts, and tumor cells, in conjunction with vascular density and perivasculature immune profiles, could significantly enrich for long-term survivors. Furthermore, relative spatial colocalization of SMA^hi fibroblasts and tumor cells compared to B cells correlated significantly with favorable clinical outcomes. Using a deep learning model trained on engineered spatial data, we can predict with high accuracy (mean AUC of 5-fold cross-validation = 0.71) how a separate cohort of patients in the NeoTRIP clinical trial will respond to treatment based on baseline TME features. These data reinforce that the TME architecture is structured in cellular compositions, spatial organizations, vasculature biology, and molecular profiles, and suggest novel imaging-based biomarkers for treatment development in the context of TNBC.</p>","PeriodicalId":18706,"journal":{"name":"Modern Pathology","volume":" ","pages":"100652"},"PeriodicalIF":7.1,"publicationDate":"2024-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142623883","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}