30-Second Takeaway
- Genome-wide copy-number profiling is maturing into a practical tool for clonality assessment and myeloma risk stratification.
- Single-assay NGS approaches can approximate or surpass conventional cytogenetics in hematologic neoplasms.
- Molecularly defined renal, prostate, and testicular tumors require pattern-based triage with focused immunohistochemistry where sequencing is limited.
- Explainable image-based AI is approaching clinically useful performance in myelodysplastic neoplasm diagnosis.
- Liquid biopsy assays in genitourinary cancers show promising accuracy but still need robust prospective validation.
Week ending December 13, 2025
Molecular and genomic tools reshaping diagnostic and prognostic practice across solid tumors and hematologic neoplasms
Genome-wide copy-number profiling reliably resolves clonality of multiple lung tumors when mutation panels are equivocal
This study evaluated genome-wide copy number aberrations (CNAs) to determine clonality in multiple cancers with pulmonary involvement when mutation panel results were ambiguous. TRACERx and routine practice cohorts included clonal and nonclonal tumor pairs, with whole-exome sequencing (WES) mutation clonality as the reference standard. All tumor pairs definitively called clonal or nonclonal by mutation analysis were concordant between WES and CNA-based clonality assessment. Among tumor pairs labeled probable nonclonal or inconclusive by mutation analysis, CNA profiling correctly reclassified almost all cases, leaving only one inconclusive pair per cohort. The authors propose a CNA-based clonality workflow for molecular diagnostics, supporting integration of genome-wide CNAs into routine reporting of multiple lung tumors.
Explainable image-based AI distinguishes MDS, AML, and normal marrow smears with high accuracy
Investigators trained end-to-end deep learning models on bone marrow smears to classify myelodysplastic neoplasms, acute myeloid leukemia, and donor marrows. The models achieved high diagnostic accuracy on internal and external validation cohorts, despite using only slide-level labels without cell-by-cell annotation. Occlusion sensitivity mapping showed that nuclear structures beyond canonical dysplastic features contributed substantially to model decisions. These findings suggest that explainable AI can capture subtle morphologic patterns in MDS not consistently recognized in manual assessment.
Practical diagnostic approach to molecularly defined renal cell carcinomas for routine surgical pathology
This review summarizes molecularly defined renal carcinomas, including TFE3/TFEB-rearranged, FH-deficient, SDH-deficient, SMARCB1-deficient renal medullary carcinoma, ALK-rearranged, and ELOC-mutated RCC. Although these tumors represent only a small fraction of RCC, they have distinctive biology, diagnostic pitfalls, and therapeutic implications. Because many laboratories lack ready access to FISH or NGS, the authors emphasize morphologic clues and immunohistochemical surrogates for initial recognition. They propose rational triage to ancillary molecular testing when available, enabling accurate classification of these rare but clinically important entities.
References
Numbered in order of appearance. Click any reference to view details.
Additional Reads
Optional additional studies from this edition.