{"title":"Retrieval-augmented generation versus document-grounded generation: a key distinction in large language models","authors":"Shunsuke Koga, Daisuke Ono, Amrom Obstfeld","doi":"10.1002/2056-4538.70014","DOIUrl":"10.1002/2056-4538.70014","url":null,"abstract":"<p>We read with great interest the article by Hewitt <i>et al</i>., ‘Large language models as a diagnostic support tool in neuropathology’ [<span>1</span>]. The authors effectively applied large language models (LLMs) to interpreting the WHO classification of central nervous system tumors; however, we wish to address a technical aspect of their study that warrants clarification.</p><p>The authors described their approach as retrieval-augmented generation (RAG). Based on the methods described, the study involved attaching a Word document containing the WHO diagnostic criteria to the prompt to guide its responses. We believe that this approach is more accurately described as document-grounded generation rather than true RAG. Document-grounded generation refers to methods where the model generates outputs explicitly based on a preprovided document, which serves as a static reference [<span>2</span>]. Unlike RAG, which retrieves information dynamically from external sources [<span>3</span>], document-grounded generation relies entirely on data embedded in the input prompt at the time of execution. In this study, the WHO criteria were provided with the prompt, which allowed the model to use this information without real-time retrieval. This method is a type of in-context learning, relying on curated contextual data embedded in the input [<span>4</span>].</p><p>Our own work provides an example of in-context learning in a different domain, namely image classification. We evaluated GPT-4 Vision (GPT-4V) for classifying histopathological images stained with tau immunohistochemistry, including neuritic plaques, astrocytic plaques, and tufted astrocytes [<span>5</span>]. Although GPT-4V initially struggled, few-shot learning with annotated examples, which is a specific application of in-context learning, significantly improved its accuracy, matching that of a convolutional neural network model trained on a larger dataset. These findings demonstrate the utility of in-context learning for both text-based and image-based tasks, with the latter presenting unique challenges for LLMs [<span>6</span>].</p><p>Although in-context learning is an effective approach, it has limitations worth considering. Since this method uses static data that are preloaded data into the prompt, errors can occur if the information is outdated or inaccurate. In-context learning may also lead to overfitting to the given context, limiting the model's ability to generalize to other scenarios. If the contextual data are overly complex, the model might misinterpret the information or fail to generate accurate outputs [<span>7</span>]. To ensure reliability, it is important to carefully select the input data, update it regularly, and consider these limitations when designing tasks.</p><p>In summary, clarifying the differences between RAG, document-grounded generation, and in-context learning is essential, especially for readers less familiar with these concepts. Nonetheless, we support the au","PeriodicalId":48612,"journal":{"name":"Journal of Pathology Clinical Research","volume":"11 1","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-01-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11736412/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143014454","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Csilla Olah, Oleksandr Shmorhun, Gilbert Georg Klamminger, Josefine Rawitzer, Lara Sichward, Boris Hadaschik, Mulham Al-Nader, Ulrich Krafft, Christian Niedworok, Melinda Váradi, Peter Nyirady, Andras Kiss, Eszter Szekely, Henning Reis, Tibor Szarvas
{"title":"Immunohistochemistry-based molecular subtypes of urothelial carcinoma derive different survival benefit from platinum chemotherapy","authors":"Csilla Olah, Oleksandr Shmorhun, Gilbert Georg Klamminger, Josefine Rawitzer, Lara Sichward, Boris Hadaschik, Mulham Al-Nader, Ulrich Krafft, Christian Niedworok, Melinda Váradi, Peter Nyirady, Andras Kiss, Eszter Szekely, Henning Reis, Tibor Szarvas","doi":"10.1002/2056-4538.70017","DOIUrl":"10.1002/2056-4538.70017","url":null,"abstract":"<p>Distinct molecular subtypes of muscle-invasive bladder cancer (MIBC) may show different platinum sensitivities. Currently available data were mostly generated at transcriptome level and have limited comparability to each other. We aimed to determine the platinum sensitivity of molecular subtypes by using the protein expression-based Lund Taxonomy. In addition, we assessed the tumor heterogeneity within the primary tumor and between the primary and lymph node (LN) metastatic sites. Thirteen immunohistochemical markers were stained in a tissue microarray with an overall number of 1,508 cores. Statistical evaluation was performed in 199 patients divided into three chemo-naïve MIBC cohorts: (1) pT3/4 and/or LN+ patients who received radical cystectomy without platinum treatment, (2) patients who received adjuvant chemotherapy (AC), and (3) patients who underwent palliative platinum treatment for metastatic disease or postoperative progression. Overall survival (OS) was used as the primary endpoint. Patients with the genomically unstable (GU) subtype had significantly better OS in the AC group compared to the radical cystectomy group (HR: 0.395, 95% CI: 0.205–0.795, <i>p</i> = 0.005). In contrast, no such association was observed for the basal/squamous (Ba/Sq) subtype. Intratumor heterogeneity was present in 19% of cases, with the lowest level in the Ba/Sq and GU tumors (14% each) and the highest level of 43% in small-cell/neuroendocrine-like tumors. There was greater subtype heterogeneity between primary tumors and LN metastases. In conclusion, immunohistochemistry-based Lund Taxonomy subtypes remain stable within the same primary tumor, with the GU subtype deriving the greatest OS benefit from AC. However, high tumor heterogeneity between the primary tumor and metastatic sites can impact the effectiveness of therapies.</p>","PeriodicalId":48612,"journal":{"name":"Journal of Pathology Clinical Research","volume":"11 1","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-01-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11736421/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143014479","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Natalia Gorbokon, Niklas Wößner, Viktoria Ahlburg, Henning Plage, Sebastian Hofbauer, Kira Furlano, Sarah Weinberger, Paul Giacomo Bruch, Simon Schallenberg, Florian Roßner, Sefer Elezkurtaj, Maximilian Lennartz, Niclas C Blessin, Andreas H Marx, Henrik Samtleben, Margit Fisch, Michael Rink, Marcin Slojewski, Krystian Kaczmarek, Thorsten Ecke, Tobias Klatte, Stefan Koch, Nico Adamini, Sarah Minner, Ronald Simon, Guido Sauter, Henrik Zecha, David Horst, Thorsten Schlomm, Lukas Bubendorf, Martina Kluth
{"title":"Loss of MTAP expression is strongly linked to homozygous 9p21 deletion, unfavorable tumor phenotype, and noninflamed microenvironment in urothelial bladder cancer","authors":"Natalia Gorbokon, Niklas Wößner, Viktoria Ahlburg, Henning Plage, Sebastian Hofbauer, Kira Furlano, Sarah Weinberger, Paul Giacomo Bruch, Simon Schallenberg, Florian Roßner, Sefer Elezkurtaj, Maximilian Lennartz, Niclas C Blessin, Andreas H Marx, Henrik Samtleben, Margit Fisch, Michael Rink, Marcin Slojewski, Krystian Kaczmarek, Thorsten Ecke, Tobias Klatte, Stefan Koch, Nico Adamini, Sarah Minner, Ronald Simon, Guido Sauter, Henrik Zecha, David Horst, Thorsten Schlomm, Lukas Bubendorf, Martina Kluth","doi":"10.1002/2056-4538.70012","DOIUrl":"10.1002/2056-4538.70012","url":null,"abstract":"<p>Homozygous 9p21 deletions usually result in a complete loss of S-methyl-5′-thioadenosine phosphorylase (MTAP) expression visualizable by immunohistochemistry (IHC). MTAP deficiency has been proposed as a marker for predicting targeted treatment response. A tissue microarray including 2,710 urothelial bladder carcinomas were analyzed for 9p21 deletion by fluorescence <i>in situ</i> hybridization and MTAP expression by IHC. Data were compared with data on tumor phenotype, patient survival, intratumoral lymphocyte subsets, and PD-L1 expression. The 9p21 deletion rate increased from pTaG2 low (9.2% homozygous, 25.8% heterozygous) to pTaG2 high (32.6%, 20.9%; <i>p</i> < 0.0001) but was slightly lower in pTaG3 (16.7%, 16.7%) tumors. In pT2–4 carcinomas, 23.3% homozygous and 17.9% heterozygous deletions were found, and deletions were tied to advanced pT (<i>p</i> = 0.0014) and poor overall survival (<i>p</i> = 0.0461). Complete MTAP loss was seen in 98.4% of homozygous deleted while only 1.6% of MTAP negative tumors had retained 9p21 copies (<i>p</i> < 0.0001). MTAP loss was linked to advanced stage and poor overall survival in pT2–4 carcinomas (<i>p</i> < 0.05 each). The relationship between 9p21 deletions/MTAP loss and poor patient prognosis was independent of pT and pN (<i>p</i> < 0.05 each). The 9p21 deletions were associated with a noninflamed microenvironment (<i>p</i> < 0.05). Complete MTAP loss is strongly tied to homozygous 9p21 deletion, aggressive disease, and noninflamed microenvironment. Drugs targeting MTAP-deficiency may be useful in urothelial bladder carcinoma. MTAP IHC is a near perfect surrogate for MTAP deficiency in this tumor type.</p>","PeriodicalId":48612,"journal":{"name":"Journal of Pathology Clinical Research","volume":"11 1","pages":""},"PeriodicalIF":3.4,"publicationDate":"2024-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11638363/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142819680","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Ethar Alzaid, Gabriele Pergola, Harriet Evans, David Snead, Fayyaz Minhas
{"title":"Large multimodal model-based standardisation of pathology reports with confidence and its prognostic significance","authors":"Ethar Alzaid, Gabriele Pergola, Harriet Evans, David Snead, Fayyaz Minhas","doi":"10.1002/2056-4538.70010","DOIUrl":"10.1002/2056-4538.70010","url":null,"abstract":"<p>Despite the existence of established standards and guidelines for pathology reporting, many pathology reports are still written in unstructured free text. Extracting information from these reports and formatting it according to a standard is crucial for consistent interpretation. Automated information extraction from unstructured pathology reports is a challenging task, as it requires accurately interpreting medical terminologies and context-dependent details. In this work, we present a practical approach for automatically extracting information from unstructured pathology reports or scanned paper reports utilising a large multimodal model. This framework uses context-aware prompting strategies to extract values of individual fields, such as grade, size, etc. from pathology reports. A unique feature of the proposed approach is that it assigns a confidence value indicating the correctness of the model's extraction for each field and generates a structured report in line with national pathology guidelines in human and machine-readable formats. We have analysed the extraction performance in terms of accuracy and kappa scores, and the quality of the confidence scores assigned by the model. We have also evaluated the prognostic value of the extracted fields and feature embeddings of the raw text. Results showed that the model can accurately extract information with an accuracy and kappa score up to 0.99 and 0.98, respectively. Our results indicate that confidence scores are an effective indicator of the correctness of the extracted information achieving an area under the receiver operating characteristic curve up to 0.93 thus enabling automatic flagging of extraction errors. Our analysis further reveals that, as expected, information extracted from pathology reports is highly prognostically relevant. The framework demo is available at: https://labieb.dcs.warwick.ac.uk/. Information extracted from pathology reports of colorectal cancer cases in the cancer genome atlas using the proposed approach and its code are available at: https://github.com/EtharZaid/Labieb.</p>","PeriodicalId":48612,"journal":{"name":"Journal of Pathology Clinical Research","volume":"10 6","pages":""},"PeriodicalIF":3.4,"publicationDate":"2024-11-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/2056-4538.70010","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142639939","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"High chromosomal instability is associated with higher 10-year risks of recurrence for hormone receptor-positive, human epidermal growth factor receptor 2-negative breast cancer patients: clinical evidence from a large-scale, multiple-site, retrospective study","authors":"Yu-Yang Liao, Jianfei Fu, Xiang Lu, Ziliang Qian, Yang Yu, Liang Zhu, Jia-Ni Pan, Pu-Chun Li, Qiao-Yan Zhu, Xiaolin Li, Wenyong Sun, Xiao-Jia Wang, Wen-Ming Cao","doi":"10.1002/2056-4538.70011","DOIUrl":"10.1002/2056-4538.70011","url":null,"abstract":"<p>Long-term survival varies among hormone receptor-positive (HR+) and human epidermal growth factor receptor 2-negative (HER2−) breast cancer patients and is seriously impaired by metastasis. Chromosomal instability (CIN) was one of the key drivers of breast cancer metastasis. Here we evaluate CIN and 10-year invasive disease-free survival (iDFS) and overall survival (OS) in HR+/HER2−– breast cancer. In this large-scale, multiple-site, retrospective study, 354 HR+/HER2− breast cancer patients were recruited. Of these, 204 patients were used for internal training, 70 for external validation, and 80 for cross-validation. All medical records were carefully reviewed to obtain the disease recurrence information. Formalin-fixed paraffin-embedded tissue samples were collected, followed by low-pass whole-genome sequencing with a median genome coverage of 1.86X using minimal 1 ng DNA input. CIN was then assessed using a customized bioinformatics workflow. Three or more instances of CIN per sample was defined as high CIN and the frequency was 42.2% (86/204) in the internal cohort. High CIN correlated significantly with increased lymph node metastasis, vascular invasion, progesterone receptor negative status, HER2 low, worse pathological type, and performed as an independent prognostic factor for HR+/− breast cancer. Patients with high CIN had shorter iDFS and OS than those with low CIN [10-year iDFS 11.1% versus 82.2%, hazard ratio (HR) = 11.12, <i>p</i> < 0.01; 10-year OS 45.7% versus 94.3%, HR = 14.17, <i>p</i> < 0.01]. These findings were validated in two external cohorts with 70 breast cancer patients. Moreover, high CIN could predict the prognosis more accurately than Adjuvant! Online score (10-year iDFS 11.1% versus 48.6%, HR = 2.71, <i>p</i> < 0.01). Cross-validation analysis found that high consistency (83.8%) was observed between CIN and MammaPrint score, while only 45% between CIN and Adjuvant! Online score. In conclusion, high CIN is an independent prognostic indicator for HR+/HER2− breast cancer with shorter iDFS and OS and holds promise for predicting recurrence and metastasis.</p>","PeriodicalId":48612,"journal":{"name":"Journal of Pathology Clinical Research","volume":"10 6","pages":""},"PeriodicalIF":3.4,"publicationDate":"2024-11-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/2056-4538.70011","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142639935","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Merijn CF Mulders, Anna Vera D Verschuur, Quido G de Lussanet de la Sablonière, Eva Maria Roes, Christoph Geisenberger, Lodewijk AA Brosens, Wouter W de Herder, Marie-Louise F van Velthuysen, Johannes Hofland
{"title":"Clinicopathological and epigenetic differences between primary neuroendocrine tumors and neuroendocrine metastases in the ovary","authors":"Merijn CF Mulders, Anna Vera D Verschuur, Quido G de Lussanet de la Sablonière, Eva Maria Roes, Christoph Geisenberger, Lodewijk AA Brosens, Wouter W de Herder, Marie-Louise F van Velthuysen, Johannes Hofland","doi":"10.1002/2056-4538.70000","DOIUrl":"10.1002/2056-4538.70000","url":null,"abstract":"<p>Currently, the available literature provides insufficient support to differentiate between primary ovarian neuroendocrine tumors (PON) and neuroendocrine ovarian metastases (NOM) in patients. For this reason, patients with a well-differentiated ovarian neuroendocrine tumor (NET) were identified through electronic patient records and a nationwide search between 1991 and 2023. Clinical characteristics were collected from electronic patient files. This resulted in the inclusion of 71 patients with NOM and 17 patients with PON. Histologic material was stained for Ki67, SSTR2a, CDX2, PAX8, TTF1, SATB2, ISLET1, OTP, PDX1, and ARX. DNA methylation analysis was performed on a subset of cases. All PON were unilateral and nine were found within a teratoma (PON-T+). A total of 78% of NOM were bilateral, and none were associated with a teratoma. PON without teratomous components (PON-T−) displayed a similar insular growth pattern and immunohistochemistry as NOM (<i>p</i> > 0.05). When compared with PON-T+, PON-T− more frequently displayed ISLET1 positivity and were larger, and patients were older at diagnosis (<i>p</i> < 0.05). Unsupervised analysis of DNA methylation profiles from tumors of ovarian (<i>n</i> = 16), pancreatic (<i>n</i> = 22), ileal (<i>n</i> = 10), and rectal (<i>n</i> = 7) origin revealed that four of five PON-T− clustered together with NOM and ileal NET, whereas four of five PON-T+ grouped with rectum NET. In conclusion, unilateral ovarian NET within a teratoma should be treated as a PON. Ovarian NET localizations without teratomous components have a molecular profile analogous to midgut NET metastases. For these patients, a thorough review of imaging should be performed to identify a possible undetected midgut NET and a corresponding follow-up strategy may be recommended.</p>","PeriodicalId":48612,"journal":{"name":"Journal of Pathology Clinical Research","volume":"10 6","pages":""},"PeriodicalIF":3.4,"publicationDate":"2024-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11544441/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142606935","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Katherine J Hewitt, Isabella C Wiest, Zunamys I Carrero, Laura Bejan, Thomas O Millner, Sebastian Brandner, Jakob Nikolas Kather
{"title":"Large language models as a diagnostic support tool in neuropathology","authors":"Katherine J Hewitt, Isabella C Wiest, Zunamys I Carrero, Laura Bejan, Thomas O Millner, Sebastian Brandner, Jakob Nikolas Kather","doi":"10.1002/2056-4538.70009","DOIUrl":"10.1002/2056-4538.70009","url":null,"abstract":"<p>The WHO guidelines for classifying central nervous system (CNS) tumours are changing considerably with each release. The classification of CNS tumours is uniquely complex among most other solid tumours as it incorporates not just morphology, but also genetic and epigenetic features. Keeping current with these changes across medical fields can be challenging, even for clinical specialists. Large language models (LLMs) have demonstrated their ability to parse and process complex medical text, but their utility in neuro-oncology has not been systematically tested. We hypothesised that LLMs can effectively diagnose neuro-oncology cases from free-text histopathology reports according to the latest WHO guidelines. To test this hypothesis, we evaluated the performance of ChatGPT-4o, Claude-3.5-sonnet, and Llama3 across 30 challenging neuropathology cases, which each presented a complex mix of morphological and genetic information relevant to the diagnosis. Furthermore, we integrated these models with the latest WHO guidelines through Retrieval-Augmented Generation (RAG) and again assessed their diagnostic accuracy. Our data show that LLMs equipped with RAG, but not without RAG, can accurately diagnose the neuropathological tumour subtype in 90% of the tested cases. This study lays the groundwork for a new generation of computational tools that can assist neuropathologists in their daily reporting practice.</p>","PeriodicalId":48612,"journal":{"name":"Journal of Pathology Clinical Research","volume":"10 6","pages":""},"PeriodicalIF":3.4,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11540532/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142590746","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Homologous recombination deficiency score is an independent prognostic factor in esophageal squamous cell carcinoma","authors":"Yulu Wang, Bowen Ding, Yunlan Tao, Lingli Huang, Qian Zhu, Chengying Gao, Mingli Feng, Yuchen Han","doi":"10.1002/2056-4538.70007","DOIUrl":"10.1002/2056-4538.70007","url":null,"abstract":"<p>Homologous recombination deficiency (HRD) represents an impairment in the homologous recombination repair (HRR) pathway, crucial for repairing DNA double-strand breaks and contributing to genomic instability in cancer. The HRD score may be a more reliable biomarker than HRR-related gene mutations for identifying patients sensitive to poly(ADP-ribose) polymerase inhibitors. Despite its relevance in various cancers, the HRD score remains underexplored in esophageal squamous cell carcinoma (ESCC). We retrospectively analyzed HRD scores in 96 ESCC patients, examining correlations with clinical characteristics and survival outcomes, and validated our findings using the TCGA dataset. Genomic sequencing utilized a custom superHRD next-generation sequencing panel, and HRD scores were calculated from 54,000 single-nucleotide polymorphisms using Kruskal–Wallis rank-sum tests and two cut-off points for analysis. Higher HRD scores correlated with advanced tumor stages, recurrence, and mutations in <i>TP53</i> and <i>ABCB1</i>, while <i>APC</i> mutations were linked to lower HRD scores. Patients with high HRD scores had significantly shorter disease-free survival (<i>p</i> = 0.013) and a trend toward shorter overall survival (OS) (<i>p</i> = 0.005), particularly those not receiving adjuvant therapy. Conversely, HRD-high patients undergoing adjuvant therapy showed a trend toward longer OS (<i>p</i> = 0.015). Multivariate analysis identified HRD as an independent prognostic factor (hazard ratio = 2.814 for recurrence, <i>p</i> = 0.015). Validation with the TCGA dataset supported these findings. This study highlights the associations between HRD scores, clinical characteristics, and genomic mutations in ESCC, suggesting HRD as a potential prognostic biomarker. HRD assessment may aid in patient stratification and personalized treatment strategies, warranting further investigation to validate the therapeutic implications of HRD scores in ESCC.</p>","PeriodicalId":48612,"journal":{"name":"Journal of Pathology Clinical Research","volume":"10 6","pages":""},"PeriodicalIF":3.4,"publicationDate":"2024-10-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/2056-4538.70007","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142523410","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Nikolas Stathonikos, Marc Aubreville, Sjoerd de Vries, Frauke Wilm, Christof A Bertram, Mitko Veta, Paul J van Diest
{"title":"Breast cancer survival prediction using an automated mitosis detection pipeline","authors":"Nikolas Stathonikos, Marc Aubreville, Sjoerd de Vries, Frauke Wilm, Christof A Bertram, Mitko Veta, Paul J van Diest","doi":"10.1002/2056-4538.70008","DOIUrl":"10.1002/2056-4538.70008","url":null,"abstract":"<p>Mitotic count (MC) is the most common measure to assess tumor proliferation in breast cancer patients and is highly predictive of patient outcomes. It is, however, subject to inter- and intraobserver variation and reproducibility challenges that may hamper its clinical utility. In past studies, artificial intelligence (AI)-supported MC has been shown to correlate well with traditional MC on glass slides. Considering the potential of AI to improve reproducibility of MC between pathologists, we undertook the next validation step by evaluating the prognostic value of a fully automatic method to detect and count mitoses on whole slide images using a deep learning model. The model was developed in the context of the Mitosis Domain Generalization Challenge 2021 (MIDOG21) grand challenge and was expanded by a novel automatic area selector method to find the optimal mitotic hotspot and calculate the MC per 2 mm<sup>2</sup>. We employed this method on a breast cancer cohort with long-term follow-up from the University Medical Centre Utrecht (<i>N</i> = 912) and compared predictive values for overall survival of AI-based MC and light-microscopic MC, previously assessed during routine diagnostics. The MIDOG21 model was prognostically comparable to the original MC from the pathology report in uni- and multivariate survival analysis. In conclusion, a fully automated MC AI algorithm was validated in a large cohort of breast cancer with regard to retained prognostic value compared with traditional light-microscopic MC.</p>","PeriodicalId":48612,"journal":{"name":"Journal of Pathology Clinical Research","volume":"10 6","pages":""},"PeriodicalIF":3.4,"publicationDate":"2024-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/2056-4538.70008","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142510889","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Joep MA Bogaerts, Miranda P Steenbeek, John-Melle Bokhorst, Majke HD van Bommel, Luca Abete, Francesca Addante, Mariel Brinkhuis, Alicja Chrzan, Fleur Cordier, Mojgan Devouassoux-Shisheboran, Juan Fernández-Pérez, Anna Fischer, C Blake Gilks, Angela Guerriero, Marta Jaconi, Tony G Kleijn, Loes Kooreman, Spencer Martin, Jakob Milla, Nadine Narducci, Chara Ntala, Vinita Parkash, Christophe de Pauw, Joseph T Rabban, Lucia Rijstenberg, Robert Rottscholl, Annette Staebler, Koen Van de Vijver, Gian Franco Zannoni, Monica van Zanten, AI-STIC Study Group, Joanne A de Hullu, Michiel Simons, Jeroen AWM van der Laak
{"title":"Assessing the impact of deep-learning assistance on the histopathological diagnosis of serous tubal intraepithelial carcinoma (STIC) in fallopian tubes","authors":"Joep MA Bogaerts, Miranda P Steenbeek, John-Melle Bokhorst, Majke HD van Bommel, Luca Abete, Francesca Addante, Mariel Brinkhuis, Alicja Chrzan, Fleur Cordier, Mojgan Devouassoux-Shisheboran, Juan Fernández-Pérez, Anna Fischer, C Blake Gilks, Angela Guerriero, Marta Jaconi, Tony G Kleijn, Loes Kooreman, Spencer Martin, Jakob Milla, Nadine Narducci, Chara Ntala, Vinita Parkash, Christophe de Pauw, Joseph T Rabban, Lucia Rijstenberg, Robert Rottscholl, Annette Staebler, Koen Van de Vijver, Gian Franco Zannoni, Monica van Zanten, AI-STIC Study Group, Joanne A de Hullu, Michiel Simons, Jeroen AWM van der Laak","doi":"10.1002/2056-4538.70006","DOIUrl":"10.1002/2056-4538.70006","url":null,"abstract":"<p>In recent years, it has become clear that artificial intelligence (AI) models can achieve high accuracy in specific pathology-related tasks. An example is our deep-learning model, designed to automatically detect serous tubal intraepithelial carcinoma (STIC), the precursor lesion to high-grade serous ovarian carcinoma, found in the fallopian tube. However, the standalone performance of a model is insufficient to determine its value in the diagnostic setting. To evaluate the impact of the use of this model on pathologists' performance, we set up a fully crossed multireader, multicase study, in which 26 participants, from 11 countries, reviewed 100 digitalized H&E-stained slides of fallopian tubes (30 cases/70 controls) with and without AI assistance, with a washout period between the sessions. We evaluated the effect of the deep-learning model on accuracy, slide review time and (subjectively perceived) diagnostic certainty, using mixed-models analysis. With AI assistance, we found a significant increase in accuracy (<i>p</i> < 0.01) whereby the average sensitivity increased from 82% to 93%. Further, there was a significant 44 s (32%) reduction in slide review time (<i>p</i> < 0.01). The level of certainty that the participants felt versus their own assessment also significantly increased, by 0.24 on a 10-point scale (<i>p</i> < 0.01). In conclusion, we found that, in a diverse group of pathologists and pathology residents, AI support resulted in a significant improvement in the accuracy of STIC diagnosis and was coupled with a substantial reduction in slide review time. This model has the potential to provide meaningful support to pathologists in the diagnosis of STIC, ultimately streamlining and optimizing the overall diagnostic process.</p>","PeriodicalId":48612,"journal":{"name":"Journal of Pathology Clinical Research","volume":"10 6","pages":""},"PeriodicalIF":3.4,"publicationDate":"2024-10-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11496567/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142510888","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}