Tumor Morphology and Genetic Correlations

🧬 What if cell shape wasn’t just a clue… but a piece of evidence we can finally quantify? As pathologists, we've spent decades training our eyes to detect subtle patterns—the texture of the cytoplasm, nuclear contours, broken symmetries. We know that shape means something. But until now, that knowledge was empirical—intuitive, not validated. 📌 A recent study in Nature Communications introduces MorphLink, a tool that takes the next step: linking clearly interpretable morphological features with spatial omics data (Pentimalli et al., 2025). 🔍 What does MorphLink do? Extracts human-readable traits (like elongation, nuclear orientation, tissue density) from H&E images. Correlates them with gene or protein expression patterns using a robust spatial index (CPSI). Reveals previously hidden structure-function relationships in the tumor microenvironment. 📊 Examples: In breast and bladder tumors, MorphLink associated nuclear orientation with CD74 and MKI67 expression—proliferation and immune signals disguised as morphology. In mouse brain tissue, it captured layered neuronal organization and linked it to genes like NRN1. In zebrafish melanoma, it outperformed deep learning models in detecting meaningful patterns—even under image blur. 💡 This isn’t about discovering new shapes. It’s about validating what we already suspected—and shifting towards a more morphological pathology, where every subtlety adds diagnostic context and molecular insight. A pathology where morphology is no longer a visual companion, but an active biomarker—interpretable, robust, and scalable. 🧠 What if the future of diagnosis isn’t about replacing the human eye… but giving it stronger arguments? 👇 Can you imagine a tool like this embedded in your daily workflow? #DigitalPathology #SpatialOmics #QuantitativeMorphology #AIinHealthcare #Biomarkers #PrecisionOncology #DigitalTransformation

Recent advances in machine learning inform precision medicine and translational research. We introduce a pioneering approach that integrates pathology imaging with transcriptomics and proteomics to identify predictive histology features associated with critical clinical outcomes in cancer. We utilize 2,755 H&E-stained histopathological slides from 657 patients across 6 cancer types from CPTAC. Our models effectively recapitulate distinctions readily made by human pathologists: tumor vs. normal (AUROC = 0.995) and tissue-of-origin (AUROC = 0.979). We further investigate predictive power on tasks not normally performed from H&E alone, including TP53 prediction and pathologic stage. Importantly, we describe predictive morphologies not previously utilized in a clinical setting. The incorporation of transcriptomics and proteomics identifies pathway-level signatures and cellular processes driving predictive histology features. Model generalizability and interpretability is confirmed using TCGA. We propose a classification system for these tasks, and suggest potential clinical applications for this integrated human and machine learning approach. A publicly available web-based platform implements these models. Interesting paper from the Clinical Proteomic Tumor Analysis Consortium: https://lnkd.in/eAArJwDv

New work in Nature reveals that the order in which cancer-driving mutations arise fundamentally shapes whether tumors can form. Using elegant mouse-model experiments, Lourenço et al. show that potent tumor-initiating mutations (such as Apc) are often eliminated by strong negative selection unless they occur within a “permissive” tissue environment created by earlier mutations. These priming events alter evolutionary trajectories, influence subsequent mutation fitness, and open pathways to malignant transformation. Analysis of ~17,000 human colorectal cancers supports these findings, highlighting interactions between KRAS and APC mutations that mirror the priming effects seen in vivo. This study is an important reminder that cancer evolution is not solely about which mutations occur, but when and in what context they arise. By demonstrating that pre-existing mutation landscapes can enable or suppress future oncogenic events, the work deepens our understanding of tumor initiation and points toward prevention strategies that target early priming states. It also underscores the need for temporally resolved genomic studies in humans to map these trajectories and identify vulnerabilities before malignant transformation begins.

Developmental mosaicism defines a distinct mechanism of genetic predisposition to multiple EGFR-mutant primary tumors. Let's start with a case: a patient with early-stage lung cancer has two EGFR-mutant lung tumors in two distinct anatomic location and no history of smoking (or other risk factors). Genetic testing shows a shared EGFR mutation between the tumors, suggesting both originated from a single mutated cell during development. Does this finding indicate a higher risk for more EGFR-mutant tumors? Does this finding indicate a need for close monitoring or the potential benefit of targeted EGFR therapies? To answer these questions we simply do not know enough about the patho-biological mechanisms driving these tumors. In this study we examined 10 patients with early-stage, resectable lung cancer who presented with anatomically distinct tumor all containing EGFR mutations. We analyzed the genetic lineage of these tumors using whole-exome sequencing and poly(G) repeat genotyping and found two key mechanisms driving tumor development: (1) Germline EGFR Variants: In two other patients, researchers discovered inherited (germline) EGFR mutations. These variants, when studied in the lab, showed moderately increased EGFR signaling, which likely contributed to tumor development. (2) Developmental Mosaicism: In four patients, the tumors originated from a non-germline cell that mutated during development, creating a "mosaic" of cells with unique mutations. These mutations then led to multiple tumors in the lungs, all carrying EGFR mutations. Thus, in addition to germline variants, developmental mosaicism defines a distinct mechanism of genetic predisposition to multiple EGFR-mutant primary tumors, with implications for their etiology and clinical management. The study emphasizes: the additional layer of information obtained through exome sequencing and Poly(G) repeat testing (for "clading" of tumors), the clinical importance of testing multiple lung tumor (clonal relationship), another important link of cancer to developmental biology, the power of interdisciplinary collaboration, and the importance of careful anatomic mapping during specimen collection, grossing, as well as cohort assembly link to publication: https://lnkd.in/eD2VGPwM link to pubmed: https://lnkd.in/eC-vRSfF DOI: 10.1038/s43018-024-00840-y    Great teamwork by Risa Burr, Ph.D. Ignaty Leshchiner, Christina Costantino, Martin Blohmer, Tilak Sundaresan, Justin Cha, Karsen Seeger, Sara Guay, Brian Danysh, Ira Gore, Raquel Jacobs, Kara Slowik, Filippo Utro, Kahn Rhrissorrakrai, Chaya Levovitz, Jaimie Barth, Taronish Dubash, Brian Chirn, Laxmi Parida, Lecia Sequist, Mari Mino-Kenudson, Shyamala Maheswaran, Kamila Naxerova, Gaddy Getz, and Daniel Haber. #lungcancer #NSCLC #EGFR #osimertinib #multisitecancer #cancerresearch #precisiononcology #biomarker #multipleprimaries #collaboration #exome #sequencing #NGS #cancer

🚀 Excited to share our latest publication in the European Journal of Cancer! Our 3rd in a quartet of 4 papers focused on #AI in #Head&Neck #Cancers this month. Our study, titled "Pathogenomic Fingerprinting to Identify Associations Between Tumor Morphology and Epigenetic States", explores the integration of computational pathology with epigenomic data to uncover novel insights into tumor biology of #OralCancer. 🔍 Key Highlights: 🎯 Developed a pathogenomic fingerprinting approach that links computer extracted pathomic image features with epigenetic modifications. 🎯 Demonstrated how morphological patterns via computational pathology can reflect underlying epigenetic states, offering potential for improved cancer diagnostics and personalized therapies. This interdisciplinary effort was made possible through collaboration with an exceptional team, including Shayan Monabbati, Germán Corredor, Tilak Pathak, MD, Craig Peacock, Kailin Yang, SHLOMO A KOYFMAN, Peter Scacheri, Ph.D., James Lewis Jr., Satish E. Viswanath, and Berkley Gryder. 📄 Read the full article here: https://lnkd.in/eSu6a9CX #ComputationalPathology #Epigenetics #CancerResearch #AIinOncology #PrecisionMedicine

A new study from researchers at The Johns Hopkins University School of Medicine and the Johns Hopkins Kimmel Cancer Center has revealed that #humanpapillomavirus (#HPV) can drive #tumor development in certain rare sinonasal squamous cell carcinomas (SNSCCs). ✅ Published in Nature Communications, the study is the first comprehensive #genome-wide characterization of HPV-associated and HPV-independent SNSCCs. Dr. Nyall London Jr., senior study author and associate professor of otolaryngology-head and neck surgery at The Johns Hopkins University, led the team in sequencing tumor and normal DNA from 56 patients. The team found that HPV-associated tumors were more likely to arise in the nasal cavity, while HPV-independent tumors tended to form in the maxillary sinus. The study also confirmed that HPV is more than a bystander in these cancers—it plays a direct role in #tumorbiology. ✅ Dr. Fernando Zamuner, co-first author and faculty research associate in the Department of Otolaryngology-Head and Neck Surgery, explained that the researchers uncovered distinct mutational patterns: HPV-independent tumors commonly had mutations in TP53, NOTCH1, KRAS, CDKN2A, and others, whereas HPV-associated tumors frequently had #mutations in KMT2D, FGFR3, KMT2C, GOLGA5, TET1, and ARID1B. Specific hotspot mutations like E542K/E545K in PIK3CA and S249C in FGFR3 were found only in HPV-associated tumors. These mutations, as well as the integration of viral DNA and an APOBEC mutational signature, reinforce HPV's causative role. Additionally, mutations in TP53 (in HPV-independent tumors) and KMT2D/FGFR3 (in HPV-associated tumors) were linked with worse survival outcomes. ✅ The team also identified active pathways contributing to tumor growth: the PI3K and YAP/TAZ pathways in HPV-associated SNSCCs and PI3K, RAS, and MYC pathways in HPV-independent cases. Using a newly developed HPV-associated SNSCC cell line, they found that combination treatment with alpelisib (a PI3K inhibitor) and verteporfin (a YAP/TAZ inhibitor) synergistically suppressed tumor cell proliferation. These findings not only highlight the biological distinctions between HPV-driven and non-HPV-driven SNSCCs but also point to potential targeted therapies. 📑 Fernando T. Zamuner, Sreenivasulu Gunti, Gabriel Starrett, Farhoud Faraji, Tiffany Toni, Anirudh Saraswathula, Kenny Vu, Anuj Gupta, Yan Zhang, Daniel L. Faden, Michael E. Bryan, Theresa Guo, Nicholas Rowan, Murugappan Ramanathan Jr., Andrew Lane, Carole Fakhry, Gary Gallia, Clint T. Allen, Lisa Rooper & Nyall London Jr. Molecular patterns and mechanisms of tumorigenesis in HPV-associated and HPV-independent sinonasal squamous cell carcinoma, Nature Communications (2025).  https://lnkd.in/eaqczN9D

📌 55 F. CECT abdomen shows prevertebral lobulated contour lesion anterior to L5-S1 vertebral body. Intensely enhancing lesion with multifocal intralesional cystic areas as well as coarse calcifications, indenting the right psoas muscle and right common iliac arteries (likely mesenteric carcinoid tumour clinically). CT guided core biopsy, pelvic SOL was performed- Outside HPE report: Poorly differentiated malignant neoplasm likely of mesenchymal origin. D/D- Desmoplastic small round cell tumor. Block received for extended IHC panel. 🔬 Sections exhibit extensive stromal infiltration by nests, trabeculae and rosettoid clusters of mildly pleomorphic and hyperchromatic round to ovoid cells with fibrillary eosinophilic to amphophilic cytoplasm. Occasional mitoses and spindle cell differentiation are noted. Prominent intratumoural capillary network and stromal hyaline deposits are in addition, identified. 🌈 • Tumour cell population is diffusely positive for Vimentin, Synaptophysin, Chromogranin and GATA3 with moderate to strong expression of INSM1 • S100 and SOX10 highlight few interspersed sustentacular cells • Cells are negative for panCK, CAM 5.2, CD99, NKX2.2, WT1 and CD45 • Ki67 labelling index is ~20% 🎯 Overall histomorphology and IHC findings are thus consistent with a PARAGANGLIOMA. This biopsy has thus, almost reached a conclusion. I say almost, since IHC for SDHB as a surrogate for SDH mutations is desirable in this context. All paragangliomas are considered to have metastatic potential. Approximately 30% of abdominal paragangliomas may ultimately metastasize. Correlation with PET-CT impression is thus to be recommended. ⭐ A reduction of sustentacular cells in paraganglioma may indicate a higher tumor grade. That, compounded with the elevated Ki67 index in this case, warrants a thorough work-up for metastasis. Regarding grading, that’s something which we, as Pathologists, can only guess at with current prognostication protocols. Diverse prognostication schemes exist, some clinical-biochemical and some incorporating pathological findings. To the best of my knowledge, none have been validated in large scale trials, and WHO’s stand on their usage is non-committal. Anyway, an excision would be the only way to go ahead.

In metastatic #colorectal #cancer, comprehensive #molecular profiling using next-generation sequencing (#NGS) is essential to identify driver mutations and guide personalized treatment decisions. Activating mutations in #KRAS or #NRAS predict lack of response to anti-EGFR therapy, while BRAF V600E mutation is associated with aggressive #tumor behavior and requires alternative targeted approaches. HER2 amplification represents a distinct molecular subgroup in colorectal cancer, conferring resistance to anti-EGFR therapies even in RAS/BRAF wild-type tumors, yet offering an actionable therapeutic target; #HER2 status is assessed by immunohistochemistry with confirmatory in situ hybridization or NGS when indicated. Additionally, pathogenic mutations in POLE or POLD1 define an ultra-mutated colorectal cancer phenotype characterized by a high neoantigen burden and excellent response to immunotherapy. Universal assessment of mismatch repair (MMR) proteins remains a cornerstone of colorectal cancer evaluation, as MMR-deficient tumors are predictive of benefit from immune checkpoint inhibitors and may reflect underlying sporadic or hereditary mechanisms. OncoDaily #oncodaily

Are your genes your destiny? " ...we studied the impact of germline genetic variation on clonal hematopoiesis (CH) in 731,835 individuals. We identified 22 new CH-predisposition genes, most of which predispose to CH driven by specific mutational events." "Cancer is a genetic disease stemming from a combination of inherited and acquired mutations. Much of our understanding of the influence of germline genetic factors on carcinogenesis comes from studies of tumor genomes. Tumors from germline mutation carriers may show characteristic genomic patterns described as mutational signatures, which reflect unique processes of mutagenesis.1–5 Beyond mutagenesis, germline genetic variation shapes tissue-specific mutational fitness, with clones bearing a selective advantage attaining dominance at the expense of others.6,7" "...we performed a systematic assessment of the impact of [pathogenic genetic variants] PGVs on CH. We identified several genes not previously linked to CH predisposition across diverse racial groups (NBN, PTPN11, ATR, BUB1B, CBL, DOCK8, ERCC1, ERCC2, ERCC3, ERCC4, ERCC6L2, FANCI, KIT, LIG4, LZTR1, MUTYH, NTHL1, PRDM9, RAD51D, RTEL1, SPRED1 and TGFBR1). Most of these conferred an increased risk of specific somatic events rather than CH overall, thus highlighting that germline predisposition to CH varies by somatic alteration. In addition, we identified five new candidate hematologic cancer predisposition genes: XRCC2, SLX4, MLH1, NTHL1 and POLE. Given that our replication cohorts had either short or no follow-up for hematologic cancer development, these genes require validation in future work." https://lnkd.in/e7FjZX4Q