Comparison of tumors by CPR assignment. (A) Biopsies were initially reviewed locally by LLMPP hematopathologists at each site to confirm HGBCL-NOS diagnosis and assign a morphological subtype. All biopsies then underwent CPR by a panel of LLMPP hematopathologists. Biopsies were either confirmed as HGBCL-NOS, reclassified as another lymphoma entity, or deemed insufficient for diagnosis. (B) Concordance of submitted morphological subtype and CPR diagnosis. Each ribbon in the alluvial plot represents a single biopsy, linking its morphological subtype assigned by the submitting site to the centralized review diagnosis. (C) Clinical and molecular features in tumors confirmed as HGBCL-NOS or reclassified as DLBCL-NOS. Comparisons of clinical, FISH, and IHC data were performed using the Fisher exact test, whereas refined COO and LymphGen classifications were compared by the χ2 test. No comparisons were statistically significant. (D) Comparison between the mutational landscape of tumors confirmed as HGBCL-NOS or reclassified as DLBCL-NOS. Genes included in the oncoplot were mutated in at least 10% of HGBCL-NOS samples and identified as a significantly mutated genes in HGBCL-NOS, DLBCL-NOS, or BL. Gene mutation frequencies for each group are shown in the bar plots on the right, with odds ratios and 95% confidence intervals for each gene comparison shown in the forest plots. (E) Copy number profiles of tumors confirmed as HGBCL-NOS or reclassified as DLBCL-NOS. The proportion of tumors with a copy number gain (red) or deletion (blue) is plotted across each chromosome. (F) PCA of RNA sequencing data. B-ALL, B-cell acute lymphoblastic leukemia/lymphoma; BL11q, Burkitt-like lymphoma with 11q aberration; COMP, composite; ECOG PS, Eastern Cooperative Oncology Group performance status; LDH, lactate dehydrogenase; Int, intermediate; MUM1, multiple myeloma oncogene 1; NA, not available; Unsp, unspecified.
