Bone and Soft Tissue Tumors
Conditions
Keywords
Sarcoma, Bone and Soft Tissue Tumors, WGS, WTS, Methylation, Whole genome sequencing
Brief summary
The goal of this observational study is to learn whether tumors of bone and soft tissue can be sorted into groups by their genes. Today these tumors are named mainly by how they look under a microscope. There are more than 150 names in use. The same tumor can be given different names by different doctors. The study includes people of any age with a bone or soft tissue tumor. All of them had tissue sampled at Karolinska University Hospital in Stockholm, Sweden. The main questions it aims to answer are: Can tests of a tumor's genes sort these tumors into clearer groups than the names used today? How often does a tumor's gene group differ from the name it was given at diagnosis? Can a computer program tell these groups apart from scanned pictures of the tissue? Researchers will compare the groups found by the gene tests with the diagnoses given at the time. They will also look at how each group did over the years that followed. This shows which way of sorting tumors better matches what happened to participants. Participants will not have extra visits, tests, or treatment for this study. The study does not change anyone's care. Researchers will: Use tumor tissue that was already taken as part of regular care Read the DNA and RNA in the tumor, which carry its genetic instructions Read the chemical marks on the tumor's DNA that switch genes on and off Scan the glass slides of the tissue into digital pictures Collect facts about treatment and health from medical records Participants who are newly diagnosed will also be asked for a blood sample. Blood shows which gene changes a person was born with and which ones started in the tumor. If the study finds a gene change that matters for a participant's care, the study team tells the treating doctors through normal hospital routines.
Interventions
None listed
Sponsors
Karolinska Institutet
Region Stockholm
Lund University
KTH Royal Institute of Technology
Uppsala University
Study design
Observational model
COHORT
Time perspective
OTHER
Eligibility
Sex/Gender
ALL
Healthy volunteers
No
Inclusion criteria
* Histopathologically diagnosed tumor or tumor-like lesion of bone or soft tissue, including benign, intermediate and malignant lesions * Tissue sampled, resected or reviewed at the study center * Archived diagnostic material available and adequate for the analyses of the relevant cohort: formalin-fixed paraffin-embedded tissue, fresh-frozen tissue, or diagnostic glass slides * For the prospective cohort, documented informed consent
Exclusion criteria
* No retrievable diagnostic material or diagnostic report * Documented objection by the participant to the research use of their material or data
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Discordance between molecularly defined class and the original histopathological diagnosis | Through completion of the primary analysis of the molecular cohort, up to [5] years after study start. | Proportion of tumors in the molecularly profiled cohort whose class assignment from integrated genomic, transcriptomic and methylation data, made without reference to the diagnostic label, differs from the WHO-based diagnosis in the original pathology report. Reported with a 95 percent confidence interval and accompanied by the adjusted Rand index for the full cross-classification. Only assignments above a pre-specified confidence threshold are counted as discordant; low-confidence assignments are reported separately. Discordant cases undergo blinded review by two sarcoma pathologists to separate reclassification from diagnostic error. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Number of stable molecular classes identified | Through completion of the primary analysis, up to [5] years. | Number of classes recovered by consensus clustering of the integrated data that meet pre-specified stability criteria, reported with the stability metrics used (consensus matrix concordance, proportion of ambiguous clustering, cophenetic correlation) across the tested range of cluster numbers. |
| Proportion of tumors in residual diagnostic categories that receive a defined molecular class | Through completion of the primary analysis, up to [5] years. | Among tumors originally reported as undifferentiated pleomorphic sarcoma, spindle cell sarcoma not otherwise specified, or an equivalent unclassified bone lesion, the proportion assigned above the confidence threshold to a molecular class that has a defined biological identity. |
| Accuracy of molecular class prediction from digitized slides | Through completion of model development and evaluation, up to [5] years. | Performance of machine learning models predicting molecular class membership from whole slide images, evaluated in held-out data split at the level of individual participants. Reported as balanced accuracy and macro-averaged area under the receiver operating characteristic curve, with 95 percent confidence intervals, plus per-class sensitivity and specificity. |
| Frequency of clinically actionable somatic alterations | Through completion of the primary analysis, up to [5] years. | Proportion of tumors carrying at least one somatic alteration classified as clinically actionable according to a pre-specified evidence framework, reported by molecular class and by original WHO diagnosis. |
| Frequency of pathogenic germline variants in cancer predisposition genes | Through completion of the primary analysis, up to [5] years. | Proportion of participants with a matched germline sample carrying a pathogenic or likely pathogenic variant in a pre-specified panel of cancer predisposition genes, classified according to current guidelines. |
| Overall survival by molecular class | From date of diagnosis to death from any cause, assessed up to [30] years. | Time from date of diagnosis to death from any cause, with participants alive at last contact censored at that date. Estimated by the Kaplan-Meier method for each molecular class with a sufficient number of participants, and reported with the number of participants per class. |
| Prognostic information added by molecular class beyond current diagnosis and grade | From date of diagnosis to death from any cause, assessed up to [30] years. | Comparison of Cox proportional hazards models for overall survival that include molecular class against models that include the original diagnosis and grade, using the concordance index and likelihood-based comparison, adjusted for age, anatomical site, tumor size and treatment. |
Countries
Sweden
Contacts
CONTACTFelix Haglund de Flon, MD PhD, Associate Professor
Outcome results
None listed